More support is not always better support
Should every learner receive the same explanation before working through a problem? In a randomized study, 139 primary pupils learned metric length conversion with different combinations of pretraining and worked examples. Researchers compared fully worked examples with partially completed examples and examined how prior knowledge changed their effects. Pupils with less prior knowledge benefited from support, while the combination mattered: the advantage of fully worked examples was stronger with pretraining. For pupils with higher prior knowledge, fully worked examples under pretraining had a negative conditional effect; without pretraining, that contrast was not significant. The study tested instructional materials, not an AI tutor. Its relevance to adaptive systems is a design question: can help respond to what a learner already understands, rather than simply supply more explanation?
Why it matters for leaders
In one tutoring pilot, ask how support is selected and when it fades. Compare outcomes by initial knowledge using fresh problems. Avoid turning a brief diagnostic score into a permanent ability label or assuming that longer explanations are better.
The evidence
A two-by-two randomized design crossed pretraining with example completeness. Prior knowledge was measured before instruction and analyzed as a moderator of the experimental conditions.
Who was studied
139 analyzed pupils in grades 4-6 at a public school in German-speaking Switzerland; mean age 10.51.
What the evidence cannot tell us
One topic, a short intervention and a coarse prior-knowledge measure limit transfer. Seventeen incomplete cases were excluded, almost all before randomization. No adaptive AI product was evaluated.
Read the original sources
Summary and leadership analysis from the Week ending 2026-11-06 brief. The edition’s selection notes explain source access and evidence limitations.