The tutor who said less

Recordings of tutorials found that the most effective tutors spoke for about a third of the session. The least effective spoke for two thirds, and their explanations were, by any reasonable standard, clearer.

The finding is not about clarity. It is about who is doing the work. A clear explanation delivered to a student who is listening produces understanding that belongs to the explanation rather than to the student: it can be followed in the moment and not reconstructed afterwards. The effective tutors interrupted themselves. They asked the student to predict the next step, waited through silences that averaged four seconds, and allowed a wrong answer to be completed before responding to it.

That last practice is the most counter-intuitive. Letting an error finish feels like permitting a mistake to take root, and the instinct of a competent explainer is to intervene at the first wrong turn. But an error that is completed can be examined by the person who made it, and the examination is where the correction becomes durable. An error interrupted is replaced by the tutor's reasoning, which the student has no reason to doubt and no way to own.

The four-second silence has its own literature. Teachers asked to wait after a question typically manage about one second before supplying the answer or rephrasing, and extending it to three or four changes who answers: longer pauses produce responses from students who were still assembling a thought rather than from those who had one ready.

None of this makes a tutorial more comfortable. The effective sessions were reported by students as harder and by tutors as slower, and both impressions were correct.