Misconception Probe¶

A diagnostic prompt — instantiates Formative Feedback Loop

Uses questions engineered so that each wrong answer reveals a specific misconception, turning a response directly into a named diagnosis rather than a score.

A Misconception Probe is a question — or short set — deliberately engineered so that the wrong answers are diagnostic: each distractor corresponds to a specific, known misconception, so a learner's choice reveals not just that they are wrong but which mental model is producing the error. Its defining move is designing the item backward from the misconceptions you want to detect, rather than forward from the correct answer. Where a quiz asks "do they know it?", the probe asks "if they don't, why — which wrong theory are they running?" — and hands back a named diagnosis, not a percentage.

Example¶

A physics teacher wants to know whether students still hold the intuition that a constant force produces constant speed rather than constant acceleration — a classic mechanics misconception. She doesn't ask them to state Newton's second law. She shows a puck sliding on frictionless ice after one steady push and offers four continuations: it keeps speeding up (correct), it moves at a steady speed, it slows to a stop, it stops at once. The distractors aren't random wrong numbers — "steady speed" is the fingerprint of the force-implies-velocity misconception, and "slows to a stop" is everyday-friction intuition smuggled into a frictionless world. When a third of the class picks "steady speed," she doesn't merely know they are wrong; she knows which wrong theory to confront next.

How it works¶

Items are designed backward from a catalogue of known misconceptions; each distractor is a hypothesis about a specific wrong model.
A response maps to a named misconception, so the raw signal and its diagnosis arrive together.
It targets conceptual fragility, not recall — built to catch understanding that looks fine until it's probed.
It samples narrowly and deeply — one concept, precisely — rather than broadly.

Tuning parameters¶

Distractor design — how carefully each wrong option maps to a real misconception; strong distractors diagnose, lazy ones just penalize.
Concept focus — how narrow the target; one misconception per item is cleanest, bundling several muddies which one fired.
Format — forced-choice (fast, maps cleanly to known misconceptions) versus open explanation (catches unanticipated models but needs interpretation).
Reasoning demand — whether it also asks why; requiring a reason distinguishes a lucky right answer from real understanding.
Timing — used before teaching (to find pre-existing intuitions) or after (to check they were dislodged).

When it helps, and when it misleads¶

Its strength is converting a response straight into a named cause, skipping the "they got it wrong, but why?" step and pointing teaching at the actual faulty model; validated concept inventories do exactly this at scale.^[1]

Its failure mode is blind spots: it only detects misconceptions it was designed to detect, so a student holding an unanticipated wrong model slips through clean. Forced-choice items can be pattern-matched or guessed, and a right answer for the wrong reason looks identical to real understanding unless the probe asks for reasoning. The classic misuse is treating probe results as a grade — punishing the very wrong answers engineered to be informative, which teaches students to hide their misconceptions instead of revealing them. The discipline that guards against this is to keep it low-stakes, ask for reasoning where it matters, and remember it sees only the misconceptions it was built to see.

How it implements the components¶

progress_signal — the response is the evidence that starts the loop.
gap_diagnosis — because each distractor maps to a known misconception, the response is the diagnosis; naming the wrong model is the probe's whole output.

It does not prescribe the fix or reteach (adjustment_rule, actionable_feedback — Practice Review, Coaching Check-In), make an honest-signal frame its purpose (low_stakes_frame — Low-Stakes Quiz), or close a recheck loop (recheck_cycle — Rapid Feedback Cycle).

Instantiates: Formative Feedback Loop — it is the signal-plus-diagnosis step, aimed at why a learner is wrong.
Sibling mechanisms: Low-Stakes Quiz · Formative Rubric · Adaptive Practice Set · Coaching Check-In · Draft Feedback Cycle · Exit Ticket · Peer Review Protocol · Practice Review · Progress Dashboard · Rapid Feedback Cycle

References¶

[1] A concept inventory — the best known being the Force Concept Inventory in mechanics — is a validated set of multiple-choice items whose distractors are drawn from documented student misconceptions, so the pattern of wrong answers diagnoses which intuitive models a learner still holds. A misconception probe applies the same design principle at the single-item scale. ↩