Coached Practice Session¶
Method — instantiates Cognitive Apprenticeship Modeling
The learner does the real task while an expert watches and intervenes in real time — cueing, questioning, and correcting reasoning and execution as they happen.
A Coached Practice Session puts the learner in the driver's seat on a real task while an expert watches and intervenes in the moment — cueing, questioning, correcting, and reinforcing as the work unfolds. Its defining feature is live, in-the-loop feedback on the learner's own performance: the correction lands while the learner is still holding the problem, so it attaches to a live mental state rather than a cold retrospective. This is the inverse of a demonstration — the expert is no longer performing but supervising a performance — and unlike a group review or a written record, its unit is a single learner's real attempt, with responsibility handed over piece by piece as they prove ready.
Example¶
At a contact-center helpdesk, a new agent takes live customer chats while a senior rep sits beside her, muted to the customer but not to her. Early on the coach intervenes often — "ask what changed since it last worked before you send that reset script" — catching a reasoning shortcut before it reaches the customer. As the shift goes on and the agent's diagnostic questions sharpen, the coach says less, letting a slightly-wrong turn play out because the agent catches and corrects it herself.
By the end of the week the agent handles the common cases solo, with the coach only spot-checking — and, more importantly, she has internalized the question-before-script habit, not because she was told it but because she was cued to it, in the moment, until it became hers.
How it works¶
- Learner drives real work. The task is authentic and consequential, not a canned exercise, so the cues and pressure are the real ones.
- Coach intervenes in the loop. Correction, cueing, and questioning happen during the attempt, aimed at reasoning and execution together — not saved for a debrief.
- Hand over responsibility in pieces. Start with the learner owning a slice, and widen the slice as performance and reasoning stabilize.
- Withhold on purpose. Let recoverable mistakes run so the learner practices catching and fixing them — intervene before harm, but after learning.
Tuning parameters¶
- Intervention frequency — how often the coach steps in. Frequent cues keep quality high but breed dependence; sparse cues build independence but risk letting bad habits set.
- Task stakes — how consequential the live work is. Higher stakes sharpen learning but narrow the room for productive error; lower stakes free the learner to fail safely but feel less real.
- Responsibility ramp — how fast the learner's share of the task grows. Faster builds confidence and independence; too fast leaves gaps the coach has to keep back-filling.
- Intervention target — correcting execution versus probing reasoning. Fixing the action is faster; questioning the reasoning is slower but transfers further.
When it helps, and when it misleads¶
Its strength is that feedback lands while the learner still holds the problem, on real work, so corrections attach to a live decision instead of a remembered one — and responsibility can be titrated up as readiness grows. It is the mechanism that most directly converts watching into doing.
Its failure modes cluster around the coach doing too much. Over-intervention produces a learner who performs well only while cued — dependence dressed as competence — and coaching that never thins out becomes permanent supervision, not apprenticeship. There is also the expertise-reversal effect: the close guidance that helps a novice becomes redundant, even counterproductive, once the learner has internalized the pattern, so a coach who keeps cueing a near-competent learner actively slows them.[n1] The classic misuse is coaching that quietly optimizes the current output — jumping in to keep this one task clean — at the expense of the learner's long-run independence. The discipline is to fade deliberately: intervene less as reasoning stabilizes, and treat the learner's own error-recovery as the goal, not a lapse in supervision.
How it implements the components¶
coaching_feedback— its core: targeted, in-the-moment correction, cueing, and questioning aimed at the learner's reasoning and execution together.authentic_practice_task— the learner practices on real, consequential work, so the cues and constraints are the ones that actually matter.progressive_practice_responsibility— responsibility is handed over in widening slices as the learner's performance and reasoning stabilize.
It governs a single learner's live attempt; it does not set the cross-context rule for when support is withdrawn for good (fading_rule — Apprenticeship Rotation), pool cases into a shared repertoire (case_repertoire — Case Conference), or elicit the learner's own after-the-fact account (learner_articulation — Learner Explanation Prompt).
Related¶
- Instantiates: Cognitive Apprenticeship Modeling — the coached session is the archetype's core "learner does it under guidance" mechanism.
- Sibling mechanisms: Apprenticeship Rotation · Learner Explanation Prompt · Paired Problem Solving · Simulation with Debrief · Shadowing with Debrief
Editorial Notes¶
Form Classification¶
Form family: Experiment, Test & Rehearsal
Rationale: The learner deliberately performs the real task while an expert observes, cues, questions, and corrects execution, generating readiness and skill through supervised practice, so its operative form is rehearsal.
Nearest alternative: Communication, Facilitation & Learning — Coaching interaction produces learning, but the mechanism requires the learner to practice the task under observation rather than receive content or participate in discussion alone.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Education & Pedagogy
Origin pattern: Single lineage
Present-day reach: Multi-domain
Rationale: Cognitive apprenticeship established learner-owned authentic practice under live expert coaching with progressively withdrawn support.
Related originating lineages:
- Psychology — Cognitive-load and expertise-reversal research explain contingent cueing and fading.
Review resolution: Both reviewers agree on education_pedagogy as primary. Reading the mechanism confirms that its defining operation belongs to that lineage; the final record retains psychology only as materially formative origin and keeps present-day application breadth separate from provenance.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
Easily confused with Paired Problem Solving: there, expert and learner work the problem together as near-peers trading reasoning; here the learner owns the task and the expert supervises and cues. The asymmetry — who holds the problem — is the difference.
[n1] The expertise-reversal effect — instructional guidance and scaffolding that help a novice become redundant, and can even hinder more advanced learners for whom the extra support adds load without benefit. It is the cognitive-load-theory reason coaching must fade rather than persist. ↩