Two-Stage Exams: A Small Study Leaves the Grouping Question Open

A group retake can produce an impressive mark. The question for a lecturer is what each student can explain afterwards, without the group beside them.

Stephan J. Kruger's study of 24 first-year accounting students offers a small test of that question. Its results call for care before using group marks to judge individual learning or choosing one grouping policy over another.

What the class tried

The students were on a university access programme in South Africa, with additional lecture hours and regular group work. In each of two cycles, they wrote a test alone and retook it the next day in groups of three.

The first cycle used mixed-ability groups; the second grouped students of similar ability. Groups were announced without mentioning marks. A week after each retake, students answered the same questions with changed numbers in an unannounced individual test.

The individual follow-up matters

Group retakes outscored the earlier individual tests in both cycles. On the later solo test, the mixed-group cycle moved from a mean of 52% to 60%; the same-ability cycle moved from 46% to 61%.

Only the latter within-cycle gain met the author's statistical significance threshold. That does not establish a difference between grouping methods: there was no direct test of their difference, and the follow-up means were almost the same.

Kruger interprets the pattern as indirect evidence that some students relied on stronger peers in mixed groups. Group discussions were not observed, so that explanation remains an inference.

Limits and context

This was one small class, with no group that retook the tests alone. Mixed groups always came first, and the cycles covered different tasks: an asset register and closing journal entries. Practice, task differences and order could contribute to the pattern. The study cannot identify an optimal grouping policy.

Evidence strength: limited and preliminary. This is one class of 24 students, with no control group, a fixed order of conditions and a different task in each round.