Exploring before instruction? The size of the task may matter

You give an introductory accounting class a worksheet on cost behaviour before explaining the topic. Students discuss fixed and variable costs, calculate totals and amounts per unit, and try to identify patterns. Some make useful connections. Others complete the arithmetic but struggle to explain what changes with activity and what stays constant.

Would an explanation first have helped? Or did the exploration ask students to consider too many unfamiliar relationships at once?

A study by Marci S. DeCaro, Campbell R. Bego, Angela K. Thompson and Ryan J. Patrick makes that second question worth asking. Published in Educational Psychology Review, Considering Cognitive Load When Exploring Before Instruction: The Potential Role of Element Interactivity compared exploration followed by instruction with the reverse sequence. Neither order consistently came out ahead.

The evidence concerns first-year engineering students learning about Python programming errors, not accounting students. Its value for accounting lecturers is a hypothesis about task design: the scope and guidance of an exploration activity may matter alongside its position in the lesson.

The same content, organised differently

Across three classroom cohorts at one US university, 1,183 students learned about three programming-error types. During the activity, groups of two or three modified a supplied working program to generate errors, received immediate software feedback and recorded their findings on a worksheet. Explicit instruction explained the error types, illustrated their causes and showed corrections.

Within each cohort, course sections were randomly assigned to activity-first or instruction-first sequences. Both orders used the same materials. This was not a comparison between students left to discover everything and students given an explanation: everyone received both exploration and instruction.

In the first cohort, where the activity addressed all three error types together, instruction-first students scored approximately five percentage points higher on both immediate factual and conceptual tests. A second cohort using a similar combined activity produced the same direction of results.

The third cohort used shorter activity and mini-lecture sequences, each focused on one error type. Here, exploration-first students scored 63.42% on the conceptual measure, compared with 57.01% for instruction-first students. That 6.41 percentage-point difference had a very small reported effect size. Basic-fact scores did not differ detectably.

The reversal is interesting, but tentative. Combined and segmented activities were taught in different semesters, and the segmented version also included more targeted prompts and worksheet guidance. The comparison therefore cannot establish that segmentation caused the change. The conceptual test in the third cohort contained only three multiple-choice items, with responses that did not fit together well as a single measure of understanding. This limits confidence in broad claims based on the score. A second statistical analysis provided little evidence to distinguish the two teaching orders.

Assignment by section, rather than student, and unrecorded group membership add uncertainty because students sharing sections or activity groups may have related results.

The authors propose that smaller exploration units may help students notice important features before instruction supplies the explanation. They did not directly measure that noticing process. Nor did the detailed results for the second combined-activity cohort and the segmented cohort demonstrate that segmentation reduced reported mental effort, despite stronger wording about cognitive load in the abstract.

Taken together, the findings support investigating the design of exploration, not declaring a winning sequence.

Keep the relationship visible

The authors frame activity complexity through element interactivity: the related pieces of information learners must consider together to understand something. This is not simply worksheet length or the number of topic headings. A short question can require several unfamiliar relationships to be coordinated.

Consider a proposed accounting illustration for students who can already calculate an average:

A workshop has a fixed monthly facility cost of $12,000. It produces either 1,000 or 1,500 units in a month. Both output levels are within its existing capacity, and the monthly facility cost remains unchanged. Consider this cost only.

Students could calculate the facility cost per unit at each output level, then discuss: How can the total cost remain fixed while the cost per unit changes?

The answers are $12 per unit at 1,000 units and $8 per unit at 1,500 units. The conceptual relationship is that the same total is spread across different numbers of units. “Fixed” describes the total cost within the specified activity range, not an unchanging amount per unit.

This keeps the necessary relationship intact while temporarily excluding variable costs, mixed costs and break-even calculations. It is bounded exploration, not an attempt to make students infer a whole topic from a table.

A lecturer could give this activity before a short explanation, using students’ tentative answers to introduce the distinction between total and per-unit behaviour. Alternatively, the lecturer could explain the relationship with a worked example first, then use the same activity as practice. Both are credible designs; the research gives a reason to compare them rather than assume one must be better.

In either design, students’ explanations can show whether they connect the unchanged total with the changing output and per-unit cost. The aim is not to make the task easy. It is to make the important relationship visible enough to learn.

Test the sequence, then test what lasts

A manageable local trial could replace part of an existing cost-behaviour session rather than add another activity. Compare exploration followed by explanation with explanation followed by the same exploration, keeping content, time, prompts and access to support constant. If the question is specifically about segmentation, compare combined and segmented versions within the same cohort where feasible, without simultaneously changing worksheet guidance.

Look separately at calculation accuracy and explanation. A correct division shows procedural performance on that item. An explanation of why total and per-unit costs behave differently provides additional evidence about the relationship students understand. Interest and perceived difficulty can be useful feedback, but neither substitutes for those outcomes.

The engineering study assessed students immediately after the complete teaching sequence. Students were instructed to complete the test individually, but access to notes, software or other aids was not sufficiently specified to establish unaided capability. There was no delayed assessment or test of transfer to meaningfully changed problems.

For the accounting illustration, a later individual question could change a relevant condition rather than merely substitute new numbers:

Producing 2,000 units requires additional space, increasing the workshop’s total monthly facility cost to $18,000. All other details are irrelevant to this question. A colleague predicts a facility cost of $6 per unit by dividing the original $12,000 by 2,000. Explain whether that prediction is justified and calculate the appropriate amount.

Students should recognise that the original constant-total assumption no longer applies. The appropriate calculation is $18,000 divided by 2,000, or $9 per unit. Their explanation should connect the changed capacity condition to the changed total, not simply produce another quotient.

Specify the resources permitted on this later task. If the aim is independent foundational capability, assess it without peers, AI or worked solutions. This would provide evidence about whether students retain and apply the relationship beyond the supported activity, although one item would not establish broad transfer.

The useful next step is modest: choose one relationship students routinely confuse, decide what they could productively explore, and compare the two teaching orders. Judge the result by what they can later recognise, explain and calculate, not just by how busy the room becomes.