Updated on
August 29, 2026
Habits of Discussion: The Talk Moves, and the Evidence
Technique 44 teaches learners to build on each other, not answer the teacher. The four habits, how to roll them out, and the trial that went backwards.
What are habits of discussion?
Habits of Discussion is technique 44 in Teach Like a Champion 3.0. It is a set of taught, rehearsed conventions that let a class talk to each other rather than take turns talking at the teacher: speak so everyone can hear, name the person you are building on, say plainly whether you agree or disagree, and add something rather than repeat a point already made. Lemov splits the technique in two, instilling the fundamentals first and then leading the discussion, and schools usually attempt the second half without the first.
A Year 8 class is discussing whether the Luddites had a point. Six learners speak. All six look at the teacher while they speak, and each one starts again from nothing. Nobody says "building on what Amira said". Nobody disagrees with anybody. It is a discussion in seating plan only, and from the back of the room it looks like a good lesson.
Habits of Discussion is the Teach Like a Champion technique aimed at exactly that room. It is technique 44 in the third edition, sitting in Chapter 9, Building Ratio Through Discussion, between Turn and Talk and Batch Process. It is a set of taught, rehearsed conventions that let a class talk to each other rather than take turns talking at the teacher. In that Year 8 room, for example, the fix is not a better question. It is that nobody has ever been taught how to pick up somebody else’s point and carry it.
The technique itself has never been trialled. The nearest evidence is the Education Endowment Foundation's evaluation of dialogic teaching, which ran across 76 schools and 4,958 Year 5 learners and found two additional months of progress in English and science. Four terms later, an addendum looking at the same cohort's Key Stage 2 results found nothing. The largest study of classroom dialogue since then asked the question directly, and found that what tracks with attainment is the kind of talk, not the amount of it (Howe et al., 2019). Both halves of that belong in any honest account, and this guide gives you both.
What You Need to Know
- It is two techniques wearing one name: teaching learners the fundamentals of listening and building, then leading the discussion once they have them. Schools usually attempt the second without the first.
- The habits are small and behavioural: speak so everyone can hear, name the person you are building on, say plainly whether you agree or disagree, and do not repeat a point already made.
- Teach one habit at a time and rehearse it on content the class already knows. A discussion is the wrong place to introduce a new rule.
- The evidence is for structured talk in general, not for this technique. No trial of Habits of Discussion exists. Say so when you introduce it in a staff meeting, because somebody will ask.
- The full-sentence rule needs the most thought, because it is the one that can quietly police a learner's home dialect rather than their reasoning.
- The failure mode is compliance: a room where everybody uses the stems correctly and nobody changes their mind is worse than the room you started with, because it now looks fine.
What Habits of Discussion Is, and Where It Sits in Teach Like a Champion
Habits of Discussion is a set of conventions a class is explicitly taught so that discussion can run between learners rather than through the teacher. Doug Lemov places it in the chapter on building ratio, meaning the share of the thinking done by learners rather than by the adult. It is a routine, taught and rehearsed like any other, not a lesson activity you can drop in on a Thursday.
The two halves Lemov actually describes
Teach Like a Champion's own training materials split the technique in two, and the split is the most useful thing in it. The first half is instilling the fundamentals: teaching and reinforcing the habits that let learners listen to one another and develop an idea as a group. The second half is leading the discussion: the planning done beforehand and the interventions made during it so that the class arrives somewhere specific.
Most schools attempt the second half. They plan a good question, they seat the class in a horseshoe, and they wait for the discussion Lemov describes. It does not arrive, because the fundamentals were never taught. The order is not a preference. A class that has not rehearsed how to disagree cannot be led into disagreeing well.
Why references disagree about the technique number
You will find Habits of Discussion cited as technique 42 and as technique 44, and both are right about different editions. Teach Like a Champion has been renumbered twice. In the 2021 third edition it is 44, verified against the publisher's own contents listing for the book. Older videos and blog posts carry the second edition number. If you are citing it in a policy document, name the edition.
Habits of Discussion is not Habits of Attention
These two get conflated constantly, including in school CPD packs, and they do different jobs. Habits of Attention is technique 48, in the chapter on procedures and routines, and it is about making listening visible: where learners look, how they sit, how you can tell attention from its absence. Habits of Discussion is about what learners say and how they connect it to each other.
The practical difference matters. A class can have excellent Habits of Attention and no Habits of Discussion at all. That is the silent, well-drilled room where every head turns to the speaker and nothing any speaker says is picked up by the next one.
The Habits Themselves, and Who They Belong To
The habits divide into what learners are taught to do and what the teacher does while they do it. The learner habits are small, concrete and behavioural, which is what makes them teachable. The teacher moves are named tools with specific jobs. A good deal of what circulates online under this heading belongs to neither, and it is worth being able to tell which is which.
The fundamentals learners are taught
Four habits do most of the work, and each one is a single observable behaviour rather than a disposition.
| Habit | What it looks like | The problem it solves |
|---|---|---|
| Speak so everyone can hear | Voice reaches the far corner without the teacher repeating it | The teacher stops being a human amplifier, which is what turns every exchange back through the front of the room |
| Name who you are building on | "Building on Amira's point about wages..." | Makes the connection audible, so the class can hear an argument being assembled rather than a list of answers |
| Say whether you agree or disagree | "I disagree with Tom, because the dates do not work" | Stops polite accumulation. Without it a discussion drifts to consensus without anyone testing anything |
| Do not repeat what has been said | Learners track what is already on the table and add to it | Forces listening. A learner who is rehearsing their own point cannot follow this rule |
Notice how little any of these ask about the quality of an idea. That is deliberate. They are the conditions under which ideas can be examined, and they are teachable in a week, which is more than can be said for the reasoning they make room for.
The three interventions the teacher makes
Teach Like a Champion names three moves for the second half of the technique, and they are more specific than general facilitation advice.
- Strategic Restate. The teacher repeats a learner's contribution, but chooses which part to repeat. Restating the half that moves the argument on is a way of steering without taking the floor.
- Knowledge Feeding. The teacher supplies a missing fact rather than a hint. When a discussion has stalled because the class does not know something, a question is cruelty and a fact is a gift.
- Stamping. Naming what the discussion established, out loud, before it closes. Without it the learners who were quiet leave with nothing, and the discussion only benefits the people who spoke.
Stamping is the one most often skipped and the one most worth keeping. A discussion that ends when the bell goes has taught the class that talk is decoration. It also links directly to checking for understanding: the stamp is your evidence that anything landed.
What has been added by other people
Several rules travel under this banner that are not Lemov's. The most common is the requirement to answer in a full sentence, which belongs to the broader Teach Like a Champion family, notably Format Matters, rather than to Habits of Discussion specifically. Sentence stems printed on a wall are a school-level adaptation, useful and widely used, but not part of the technique as published.
This matters for two reasons. A staff meeting goes better when you can say which parts are the published technique and which are your school's additions, because the additions are then open to change. And the full-sentence rule carries a cost that the four core habits do not, which the section on SEND and dialect below sets out. For the participation side of this, see means of participation and format matters.
What the Evidence Says About Structured Talk
There is no trial of Habits of Discussion. There is a reasonable body of evidence about structured classroom talk in general, and it says something more specific than "discussion is good": the kind of talk matters far more than the amount of it, and gains are easier to find in comprehension than in general attainment. That is a narrower claim than most CPD makes, and it is the defensible one.
The largest British trial, and what happened next
The Education Endowment Foundation trialled Robin Alexander's dialogic teaching approach in 76 schools in Leeds, Bradford and Birmingham. Of the 4,958 Year 5 learners randomised, 3,912 completed post-tests. The trial reported two additional months of progress in English and science and one in maths, with a three padlock security rating, and the same two months for learners eligible for free school meals.
Then came the part almost nobody repeats. The EEF published an addendum in April 2019 following the same cohort to their Key Stage 2 tests. Its own words: the analyses "provide no definitive evidence that Dialogic Teaching schools made additional progress, measured by attainment in KS2 national tests, when compared with controls", and "the estimated effects are very small and not statistically significant". The effect sizes were 0.03 for reading and 0.06 for maths.
Read that fairly in both directions. The addendum's own authors point out that four terms passed with no data on what schools did in between, so this is not a demonstration that the approach failed. It is a demonstration that a two month gain measured immediately did not show up in a national test eighteen months later, which is a normal and sobering thing to happen to a promising intervention. Anyone selling you talk protocols on the strength of the headline figure has read half the evidence.
It is the kind of talk, not the amount
Howe et al. (2019) asked the sceptical question directly, analysing dialogue in seventy-two Year 6 classrooms against later attainment. The relationship they found was real but conditional. Particular features, learners elaborating and querying each other's contributions, tracked with outcomes. The sheer quantity of talk did not.
That finding is the strongest argument for this technique that exists, and it is indirect. Habits of Discussion is a way of manufacturing exactly the features Howe found mattered: building on a named contribution is elaboration, and saying you disagree is querying. It is a plausible mechanism, not a tested one, and the difference should be stated plainly.
Nystrand et al. (2003) is the useful companion, because it measured how rare this is. Tracking discourse in real classrooms minute by minute, it found genuinely dialogic episodes were brief and uncommon even in lessons that looked discussion based. If your class is not doing this yet, you are the norm rather than the exception.
Discussion and comprehension
The clearest positive evidence sits in reading comprehension. Murphy et al. (2009) meta-analysed classroom discussion approaches and found that several produced large increases in learner talk, but that increases in talk did not automatically produce increases in comprehension. The approaches that lifted comprehension were the ones that pushed learners towards higher level thinking about the text.
Soter et al. (2008) went further into the discourse itself and identified which features distinguish discussions that indicate high level comprehension. Kim and Wilkinson (2019) is the honest overview: dialogic teaching is defined differently by nearly every research group using the term, which is one reason the trial evidence is harder to add up than it should be. If you want the framework with the most careful theory behind it, read our guide to accountable talk, and note that it has never been trialled either.
Teaching the Habits Without Killing the Talk
The rollout is where this technique succeeds or fails. Introduced as a list of six rules on a slide, it produces stilted, self conscious talk that everybody hates by half term. Introduced one habit at a time, on content the class already knows, it becomes invisible within about three weeks and the discussion underneath it gets better.
One habit at a time
Pick the habit your class most needs and teach only that one for a fortnight. For most secondary classes it is naming who you are building on, because it is the one that changes the shape of the conversation fastest. For most primary classes it is audibility, because without it nothing else can happen.
Rehearse it on content the class finds easy. A new rule and a hard idea at the same time is two loads at once, and the rule is the one that gets dropped. Ben Newmark's point about starters applies here too: the routine has to be a doddle before it can carry anything heavy.
Sentence stems, and when to retire them
Stems work. "I agree with ... because", "I want to build on ...", "I see it differently from ... because" give a learner a way in when the social cost of speaking is the real barrier. They are a scaffold, and the defining property of a scaffold is that it comes down.
Leave them up all year and two things happen. Fluent learners produce the stem and then say nothing inside it, which is the compliance failure in miniature. And the class learns that the form is what you are marking. Plan the removal when you plan the introduction: half a term of stems on the wall, then stems on a card for the learners who still want one, then nothing.
The first three lessons
A rollout that works looks roughly like this, and it takes about ten minutes of three lessons rather than a whole lesson of one.
| Lesson | What you do | What you are watching for |
|---|---|---|
| One | Name the habit, model it badly, then model it well. Ask the class to say what changed | That they can describe the behaviour, not that they agree it is a good idea |
| Two | Run a short paired talk on easy content, then take three contributions using the habit. Correct the form, not the idea | Learners reaching for the stem without being prompted |
| Three | Run it in a real discussion. Stamp what was established at the end | Whether anyone changed their position. If nobody did, the habit is running and the discussion is not |
Two techniques pair naturally with this. Cold calling is how you keep the habit from becoming a volunteers only routine, and wait time is what makes a considered second contribution possible at all. Three seconds of silence is where a learner assembles the sentence that names what somebody else said.
Adapting It for SEND, EAL and Home Dialects
The four core habits adapt well. The added rules, particularly the full sentence requirement, need the most thought, because they can end up assessing how a learner speaks at home rather than how well they are reasoning. This is the part of the technique where a school's decisions matter more than Lemov's, and where getting it wrong does real harm quietly.
The full-sentence rule is the one to think hardest about
A requirement to answer in a full sentence in Standard English is doing two jobs at once. It is asking a learner to make their reasoning explicit, which is defensible and useful. It is also asking them to translate out of their home dialect in public, at speed, in front of their peers, which is a different demand and carries a different cost.
The workable separation is to require the reasoning and teach the register. Insist that a learner says why, and accept it in the words they have. Then teach Standard English forms explicitly, as a thing you use in some settings, at a moment when nobody is also trying to hold an argument in their head. Correcting dialect during a discussion teaches most learners to stop volunteering.
Processing time and the learner still forming the sentence
Naming who you are building on requires holding someone else's point in working memory while composing your own. For a learner with slower processing or a language disorder, that is the whole difficulty, and the habit can make speaking harder rather than easier.
Three adjustments cost nothing. Let the learner point at the person rather than remember the name. Allow a written note before speaking, which is what everybody writes is for. And warn them in advance that you will come to them, which removes the retrieval demand without removing the contribution.
The learner who will not speak
Habits of Discussion assumes speaking. For a selectively mute learner, or one in the silent period of English acquisition, the technique as written has nothing to offer, and a well meaning insistence is actively harmful. Give them a non spoken route into the same thinking: writing a contribution for someone else to read out, or agreeing and disagreeing on a mini whiteboard while others speak.
That is not a lowered expectation. It is the same expectation on the reasoning with a different means of participation, which is exactly the distinction Teach Like a Champion draws elsewhere and does not draw here.
Limitations and Critiques
Four objections deserve an answer, and a colleague could raise any of them in a staff meeting. They concern the absence of a trial of this technique, what happened to the nearest trial that does exist, the risk that a talk protocol produces compliance instead of thinking, and the specific critique of scripted classroom moves in the schools this approach came from.
The technique itself has never been trialled
There is no randomised trial, no quasi experimental study, and no published evaluation of Habits of Discussion. Everything above rests on evidence about structured talk in general, which is a different and broader thing. That is not a reason to avoid the technique, and it is a reason not to call it evidence based without a qualifier. The same is true of most named classroom techniques, including several in our guide to all 63 Teach Like a Champion techniques.
The nearest evidence went backwards on follow-up
The dialogic teaching addendum above is the honest ceiling on what anyone can claim. A two month gain that does not appear in a national test eighteen months later may reflect schools stopping the practice, or may reflect a real but shallow effect. The trial cannot tell you which, and neither can anybody quoting it at you.
A protocol can produce agreement rather than reasoning
The failure this technique invites is a class that has learned the moves and uses them to be pleasant. Everyone names who they are building on, everyone builds, nobody disagrees, and the discussion arrives exactly where it started. It looks better than the room you began with, which is what makes it dangerous.
The diagnostic is simple and worth running termly. At the end of a discussion, ask how many learners changed their mind about anything. If the answer is habitually none, the habits are running and the discussion is not, and the fix is a better question rather than more protocol.
The critique of scripted moves
Joanne Golann's ethnography of a no excuses school, Scripting the Moves (2020), is the serious version of this criticism. Her argument is not that structure harms learners but that heavily scripted behaviour can teach compliance and deference rather than the confident self advocacy the schools intend, and that this falls hardest on the learners the approach is meant to serve.
Applied here, the test is whether your habits give learners more power in the conversation or less. A rule that makes disagreement sayable gives them more. A rule about how to sit while disagreeing gives them less. Both can appear on the same laminated poster, and only one of them is Habits of Discussion.
References and Further Reading
Every source below was checked twice: once to confirm the record exists, and again against Crossref to confirm it is the work being cited. The Education Endowment Foundation figures were read from the evaluation report and its addendum directly rather than from a summary. Links go to the publisher, never to an aggregator.
- Dialogic Teaching: Evaluation Report and Executive Summary. Jay, T., Willis, B., Thomas, P., Taylor, R., Moore, N., Burnett, C., Merchant, G., and Stevens, A. (2017). Education Endowment Foundation. The largest British trial of a structured talk approach. 76 schools, 4,958 Year 5 learners randomised, 3,912 completing post-tests. Two months' additional progress in English and science.
- Dialogic Teaching: Addendum to the Evaluation Report. Education Endowment Foundation (2019). The follow-up nobody quotes. It found "no definitive evidence" of additional progress at Key Stage 2, with effect sizes of 0.03 in reading and 0.06 in maths, neither statistically significant.
- Teacher-student dialogue during classroom teaching: Does it really impact on student outcomes?. Howe, C., Hennessy, S., Mercer, N., Vrikki, M., and Wheatley, L. (2019). Journal of the Learning Sciences, 28(4-5), 462-512. The sceptical question asked directly. The kind of talk predicted outcomes; the amount did not.
- Examining the effects of classroom discussion on students' comprehension of text: A meta-analysis. Murphy, P. K., Wilkinson, I. A. G., Soter, A. O., Hennessey, M. N., and Alexander, J. F. (2009). Journal of Educational Psychology, 101(3), 740-764. More talk did not reliably mean more comprehension. The approaches that worked pushed learners towards higher level thinking about the text.
- What the discourse tells us: Talk and indicators of high-level comprehension. Soter, A. O., Wilkinson, I. A. G., Murphy, P. K., Rudge, L., Reninger, K., and Edwards, M. (2008). International Journal of Educational Research, 47(6), 372-391. Which features of classroom discourse actually mark comprehension, rather than which ones feel productive.
- Questions in time: Investigating the structure and dynamics of unfolding classroom discourse. Nystrand, M., Wu, L. L., Gamoran, A., Zeiser, S., and Long, D. A. (2003). Discourse Processes, 35(2), 135-198. Genuinely dialogic episodes were brief and rare even in lessons built around discussion. Useful perspective before judging your own classroom.
- What is dialogic teaching? Constructing, deconstructing, and reconstructing a pedagogy of classroom talk. Kim, M.-Y., and Wilkinson, I. A. G. (2019). Learning, Culture and Social Interaction, 21, 70-86. The term means different things to different research groups, which is one reason the evidence is hard to add up.
- Scripting the Moves: Culture and Control in a "No-Excuses" Charter School. Golann, J. W. (2020). Princeton University Press. The serious version of the compliance critique, from inside the school culture these techniques came from.
- Teach Like a Champion 3.0: 63 Techniques that Put Students on the Path to College. Lemov, D. (2021). Jossey-Bass. The primary source. Habits of Discussion is technique 44, in Chapter 9. A trade book setting out practice, not a study.
