Faded Scaffolding

Faded scaffolding is a scaffold-removal pattern in which an instructor or generative AI tutor gradually withdraws support across successive attempts so the learner shoulders more of the cognitive work each pass. The fade typically moves through modeling, partial completion, learner-chosen next steps, similar problems with fewer hints, unaided strategy explanation, and transfer to a new context. Each transition is gated by evidence of success at the prior level. The fade is what makes scaffolding scaffolding — without removal, support hardens into dependency, producing high assisted performance and low unassisted competence. It is broader than worked-example fading, spanning modeling through transfer rather than only example completion.

See also

• Productive struggle — the difficulty band a fade is calibrated to preserve.
• Cognitive offloading — what a tutor that never fades produces.
• Performance–learning gap — the failure pattern that motivates the fade.