Seventy-eight per cent teacher talk

The data describes an interaction pattern, not a class trait.

• 0.8 seconds' wait time is below the threshold at which most learners can process a question and formulate an answer. Marcus is effectively answering his own questions, which is why he talks 78% of the time.
• 27 of 31 questions are closed. A closed question has one right answer and the teacher already knows it, so the learner's task is not to communicate but to be evaluated. Learners who are unsure of the answer reduce risk by staying silent. Silence here is rational, not passive.
• Six learners doing 84% of turns is what happens when the teacher takes volunteers. Volunteering selects for confidence and speed, and it lets everyone else opt out permanently.

Why points will fail:

• It rewards public answering, which is precisely the behaviour the silent learners are avoiding. It raises the cost of being wrong.
• It will mostly reward the six who already speak, widening the gap.
• It re-labels a structural problem as a motivational one, so the pattern stays intact and Marcus concludes the class is 'unmotivated' rather than that his questioning is.

Telling phrase: 'won't speak' again attributes to attitude what the transcript explains structurally.

First change: replace volunteer nomination with rehearsal plus targeted nomination, and extend wait time.

Operationally: ask the question, give 20–30 seconds of pair talk ('tell your partner, then I'll ask three of you'), then take answers from named pairs rather than hands. Extend wait time after whole-class questions to 3–5 seconds and say nothing during it.

Prediction, and I would say this to Marcus before he starts:

• The first lesson will feel worse. Three to five seconds of silence feels like a failure to a teacher, and he will want to fill it. If he fills it, nothing has changed.
• The six dominant learners will initially answer over the pairs. The fix is procedural, not personal: nominate from pairs, not from hands, and don't take a second answer from the same learner.
• Two or three of the quiet learners will produce short, low-risk answers first, often reading from what their partner said. That is the behaviour to build on, not to push past.

Success measure for lesson two: number of different learners contributing to whole-class talk, and average length of student turn - not volume of talk.

First, separate two hypotheses, because the remedy differs:

• Pattern hypothesis - they are silent because the lesson gives them no safe slot, and they have learned that not speaking goes unnoticed. Fix is the pair rehearsal above plus seating them where a partner exists.
• Individual hypothesis - something specific: a hearing or processing issue, a dispute with another learner, a language barrier not visible in class, or a bad relationship with the subject. Fix is a short private conversation, not a classroom technique.

What I would not do: put them on the spot publicly in the first lesson after the change. That converts a low-stakes pair slot into a high-stakes public failure and confirms the reason for silence.

Sequence: seat them with a patient partner; give a written or one-word role in the pair task; observe whether they speak to the partner; if they do, widen gradually; if they still do not after two or three lessons, have the private conversation. Points schemes do not distinguish these two hypotheses and so cannot resolve either.