Turn and Talk and Talk Partners: A Teacher’s Guide

The most-used talk routine in English classrooms is also the least-taught. Most teachers reach for it several times a lesson. Very few have ever been shown how to run it properly.

Turn and talk is a short, structured routine in which every learner discusses a teacher's prompt with one partner for roughly thirty to sixty seconds, one speaking while the other listens, then swapping, before the class comes back together. In British schools the same routine is usually called talk partners, and the two names describe the same thing.

That teaching gap matters, because the evidence is more mixed than the technique's popularity suggests. The Education Endowment Foundation rates oral language interventions at +6 months additional progress on extensive evidence, while its collaborative learning strand sits at +5 months on limited evidence and carries a caveat almost nobody repeats: "there are smaller positive impacts for both paired work" than for groups of three to five. The largest British trial of a structured talk approach, the Education Endowment Foundation's evaluation of dialogic teaching, ran with 76 schools and 4,958 Year 5 learners and reported +2 months in English and science. Howe et al. (2019) then asked directly whether classroom dialogue affects outcomes at all, and found the relationship real but conditional on the kind of talk rather than the amount of it. So paired talk is not free progress. It earns its place on frequency and low cost, not on effect size.

Anyone who has taught knows the failure mode. In a Year 5 maths class, told to talk to their partner about how they solved a problem, forty seconds produces "I got 24, what did you get" and then silence. The routine ran. No thinking happened.

‍

What Is Turn and Talk?

Turn and talk is a paired discussion routine lasting under a minute, used inside a lesson rather than as an activity in its own right. The teacher poses a prompt, learners turn to a named partner, one talks while the other listens, they switch, and the class comes back together. It belongs to the wider family of classroom talk routines covered in the importance of oracy in classroom learning.

‍

Turn and Talk and Talk Partners: One Routine, Two Names

American schools say turn and talk. British schools usually say talk partners, and some say talking partners or partner talk. The terminology split is almost total: search data shows British teachers looking for talk partners and American teachers looking for turn and talk, with barely any crossover. If you have arrived here from one name, the other is the same thing described by a different profession.

This matters practically in international schools, where a British-trained teacher and an American-trained teacher can describe the identical routine to each other and both believe they are talking about something new.

‍

What the Routine Actually Is

Five steps, in order. Pose the prompt. Pair learners with a named partner. Give a fixed, short window for talk. Signal the switch so the second learner gets their turn. Bring it back with a share back that someone is accountable for.

Most classroom versions drop steps four and five. That is where the routine stops working.

‍

What the Evidence Says, and What It Does Not

The honest position is that classroom talk has good evidence, paired talk specifically has thinner evidence, and the two are routinely reported as one thing. Oral language work is among the strongest entries in the Education Endowment Foundation's toolkit. Paired talk is a subset of collaborative learning, where the evidence base is smaller and the recommended group size is larger than two.

‍

‍

‍

The Education Endowment Foundation puts oral language interventions at +6 months additional progress, rated high impact, very low cost, on extensive evidence. That is one of the strongest entries in its toolkit. Its collaborative learning strand is rated at +5 months, moderate impact, very low cost, on limited evidence, and this is where the caveat lives: the evidence indicates groups of three to five are most effective, with smaller positive impacts for paired work and for groups larger than five.

Howe et al. (2019) asked the sceptical question directly in the title of their paper: does teacher-learner dialogue really affect outcomes? Working across a large corpus of lessons, the Cambridge team found the relationship between dialogue and attainment real but conditional on the kind of talk, not the amount. Howe and Abedin (2013) reviewed four decades of classroom dialogue research and reached a similar place: the field has consistently shown that talk can matter and has struggled to show that any particular arrangement of it reliably does.

Wait time sits underneath all of it. Rowe (1974) found that extending the pause after a question from under a second to around three seconds changed the length, confidence and logical structure of learner answers. Tobin (1987) reviewed the follow-up work and connected longer wait time to higher-order thinking specifically. If you take one thing from the research into a turn and talk, take the three seconds before the prompt, not the routine itself.

‍

Why Paired Talk Still Earns Its Place

Not on effect size. On frequency and cost. A turn and talk takes forty seconds, needs no resources, no grouping, no reorganisation of the room, and can be run six times in a lesson. Groups of four produce better talk and cost five minutes to form. Over a term, the routine you can run constantly may do more than the arrangement that works better and happens rarely.

That is a defensible argument. It is not the same argument as "research shows turn and talk works", and it is worth being clear which one you are making when a colleague asks.

‍

Designing a Prompt Worth Forty-Five Seconds

Prompt design is the single biggest failure point in paired talk, and it is barely discussed anywhere. A weak prompt cannot be rescued by good routine management, better timing or firmer expectations. If the question has one obvious answer, learners will swap that answer in four seconds and then talk about something else, exactly as you would in their position.

The test is simple. Ask whether two learners could disagree about the answer and both have something to say. If not, you have set up answer-swapping.

Prompt that produces answer-swapping	Prompt that produces reasoning
What did you get for question 3?	One of you got 24 and one got 18. Work out where the difference came from.
What is photosynthesis?	A plant in a dark cupboard still grows for a week. Explain to your partner how that is possible.
Did you like the poem?	Which single word in the poem is doing the most work? Convince your partner.
What are the causes of the war?	Rank these three causes. Your partner has to agree with your order or argue for a different one.

Mercer, Wegerif and Dawes (1999) drew the distinction that makes sense of this. Disputational talk is disagreement without reasoning. Cumulative talk is agreement without challenge, which is what most unstructured pairs produce. Exploratory talk, where learners engage critically but constructively with each other's ideas, is the kind that helps, and it does not appear on its own. It has to be taught. Mercer and Sams (2006) showed the same thing in maths: children taught explicitly how to talk about problems improved at solving them, and children simply put in pairs did not.

Kawalkar and Vijapurkar (2013) traced the same pattern in science through the teacher's questions rather than the learners' talk. Questions that asked for explanation and justification produced longer, more reasoned responses than questions that asked for recall, which is the same finding arriving from the other direction.

‍

‍

‍

‍

Sentence Stems That Do Real Work

"I think ___ because ___" is a floor, not a ceiling. It gets a reluctant learner started and it does nothing for a class that can already produce a sentence.

The stems worth teaching are the ones that do a job the learner cannot yet do unaided: disagreeing without falling out, building on someone else's idea rather than replacing it, and summarising what a partner said before responding to it.

• I agree with the first part, but I am not sure about ___
• Building on what you said, ___
• So you are saying ___. Have I got that right?
• What made you choose ___ rather than ___?
• I had a different answer, so one of us has gone wrong somewhere. Let us find it.

The last one is worth more than the rest combined, because it turns a disagreement into a joint problem. More on scaffolding classroom talk in talk-based learning strategies.

‍

Running the Routine: Timing, Cueing and the Switch

Thirty to sixty seconds is the working range for most prompts. Longer than that and pairs run out of prompt and drift into conversation. Shorter and the second learner never gets a turn at all. Within that window, four moves decide whether the routine produces thinking or simply produces noise: wait, name who starts, signal the switch, and stop early.

Four things make the difference between a routine and a gap in the lesson.

1. Wait first. Three seconds of silence after the prompt, before anyone turns, so learners arrive at the pair with something to say. Rowe's finding applies here more than anywhere.
2. Name who speaks first. "Partner on the left starts." Without it, the confident learner starts every time, in every pair, all year.
3. Signal the switch out loud. Half the value of the routine is in the second learner's turn, and it is the half most often lost.
4. Stop while the energy is still up. If you wait for the room to go quiet, you have waited too long, and you have taught the class that the routine ends when they get bored.

Writing before talking is worth trying with classes that produce thin talk. Thirty seconds of silent jotting, or one line on a mini whiteboard placed between the partners, gives both learners something concrete to talk about and makes the talk visible to you as you circulate.

‍

What to Do While They Are Talking

Not admin. This is the only point in the lesson where you can hear thirty learners reason in ninety seconds, and it is routinely spent tidying the desk.

Webb (2009) found the teacher's moves during and immediately after group talk to be decisive in whether the talk did anything. Gillies (2006) recorded what teachers actually say during small-group work and found a great deal of managing and comparatively little probing. Listen to three pairs, choose the misconception you will address, and plan the share back around it while they are still talking.

‍

The Share Back, and Why Hands Up Fails Here

If you take hands after a turn and talk, you hear from the learners who were already going to speak, and the routine has changed nothing about who participates.

Cold calling works better after paired talk than at any other moment, because every learner has just rehearsed an answer with a partner. The usual objection to cold calling, that it puts learners on the spot unprepared, does not apply: they have had forty seconds of preparation and a witness. Asking a learner to report their partner's idea rather than their own raises the stakes on listening as well.

‍

Choosing, Seating and Rotating Talk Partners

Pairs go stale. A learner who has sat with the same partner since September has stopped hearing anything new, and the pair has settled into a fixed division of labour where one thinks and one writes.

Rotating partners on a schedule, every two or three weeks, keeps the talk fresh and spreads the social work of the class. It also stops the quiet learner being permanently paired with the one confident partner who has learned to carry them.

The dominant-partner problem needs a structural fix, not a request. Asking a learner to talk less rarely works. Assigning roles does: Partner A explains while Partner B records, then they swap for the next prompt. The role, not the personality, decides who talks.

Mixed-attainment pairing is usually the right default, with one caution. A large gap between partners tends to produce explaining in one direction rather than talk in both, which benefits the explainer and can leave the listener passive. Pairs that are close in current understanding argue more, and arguing is the point.

‍

Where Turn and Talk Goes Wrong

Turn and talk fails in six recognisable ways, and five of them are design problems rather than behaviour problems. The routine is usually blamed when the prompt was at fault. Below is each failure mode with the fix that actually addresses it, ending with the one that matters most and gets discussed least: paired talk can rehearse a wrong answer.

1. It becomes a chat. Almost always a prompt problem rather than a behaviour problem. If there is nothing to reason about, learners will find something else to talk about.
2. The same learner speaks every time. Name who starts, and rotate it.
3. It is used before anyone has anything to say. Paired talk consolidates and challenges thinking. It cannot manufacture knowledge that has not been taught yet.
4. It runs too long. Sixty seconds is usually the ceiling. Watch the room, not the clock.
5. Nobody is accountable for listening. If the share back only ever asks for a learner's own view, listening is optional. Ask for the partner's view instead.
6. It rehearses a wrong answer. The most serious of the six, and the least discussed.

That last one deserves its own paragraph. Graham Nuthall's classroom research found that learners frequently persuade their classmates that incorrect information is true, and that the most easily persuaded are those with the weakest knowledge of the topic. A turn and talk with no check afterwards is an efficient way of distributing a misconception around the room, in the learners' own words, with a peer's authority attached. The fix is not to stop using the routine. It is to always follow it with something that surfaces what was actually said, whether that is cold calling, a hinge question or a quick exit ticket.

‍

Turn and Talk With SEND and EAL Learners

This is where most guidance on the routine stops, and where the routine most needs thought. A technique that requires every learner to speak on cue is not automatically inclusive because it involves everyone.

‍

Speech, Language and Communication Needs

Norbury et al. (2016) screened over 7,000 children and assessed 529 in depth, and put the population prevalence of language disorder at 9.92 per cent. In their words, at school entry approximately two children in every class of thirty will experience language disorder severe enough to hinder academic progress. Those two learners are in every turn and talk you run.

Practical adaptations: give the prompt in writing as well as aloud, allow longer than the standard window, permit a single word or a drawing as a contribution, and pitch the question so it can be answered at more than one level. Our guide to Blank level questions is useful for judging whether a prompt is reachable. More on identification and classroom support in our guide to speech, language and communication needs.

‍

Selective Mutism

This section is written out in full rather than summarised, because a compressed version of it is dangerous.

Selective mutism is an anxiety condition, not a choice and not defiance. A learner with selective mutism is consistently unable to speak in specific situations, typically school, while speaking normally elsewhere. The single most important thing to understand as a teacher is this: increasing the demand for speech increases the anxiety, and increased anxiety makes speech less likely, not more. Encouragement, waiting expectantly, rewarding a word, or telling the class that we are all going to hear from everyone today, all raise the demand.

For a learner with selective mutism, a compulsory turn and talk is a compulsory failure, repeated several times a lesson.

The adaptation is a genuinely equal non-verbal role, agreed with the learner in advance and never announced to the class as a concession. That might be recording the pair's thinking on a whiteboard, drawing the idea, holding up a prepared card, or writing the sentence the partner will read out. It is not a lesser task and it should not look like one. Where a learner has a plan in place, follow it rather than improvising, and coordinate with the speech and language therapist. The Royal College of Speech and Language Therapists and SMIRA, the Selective Mutism Information and Research Association, both publish current guidance for schools. Our guide to supporting a learner with selective mutism covers the classroom detail.

‍

EAL Learners and International Classrooms

Paired talk is often better for a learner using English as an additional language than whole-class discussion, because the audience is one person rather than thirty and the stakes are correspondingly lower. Two adjustments make it work harder. Pair a newer speaker with a patient, fluent partner rather than with another new arrival, and allow thinking in a first language before talking in English. Pre-teaching the two or three words the prompt depends on turns an impossible task into a reachable one.

‍

Learners Who Struggle to Sustain Attention

Short, frequent paired talk suits learners who find long stretches of listening difficult. Locke and Fuchs (1995) found peer-mediated formats improved on-task behaviour for children with behaviour disorders. The structure matters more than the duration: a named partner, a stated job and a short window give a shape to hold on to.

‍

How Turn and Talk Differs From Think, Pair, Share

They are related and they are not the same. Turn and talk is a routine: quick, informal, used many times in a lesson, and often without a formal thinking stage at all. Think, pair, share is a protocol with three named and separately timed phases, used once or twice, usually on a bigger question that deserves individual thinking time first.

The practical difference is the silent stage. If the question needs a minute of individual thought before anyone speaks, you want the longer protocol. If you want to interrupt a explanation and check the room is still with you, you want the forty-second version. Our full guide to the think, pair, share protocol covers the longer form.

‍

Limitations and Critiques

Four qualifications belong with any recommendation of this routine, and each is one a thoughtful colleague could raise in a staff meeting. They concern the strength of the evidence for pairs specifically, what happened when dialogic teaching was followed up, whether talk routines can become compliance exercises, and the risk that paired talk distributes error rather than correcting it.

The evidence for pairs specifically is weaker than for groups. The EEF's collaborative learning summary states that groups of three to five are most effective, with smaller positive impacts for paired work. Any claim that turn and talk is evidence-based should carry that sentence with it.

Dialogic teaching did not survive its own follow-up. The EEF's trial of Alexander's dialogic teaching approach, run with 76 schools and 4,958 Year 5 learners, reported +2 months in English and science and +1 month in maths. That last figure is worth reading carefully: the maths effect size was 0.09 with a confidence interval running from -0.04 to 0.20 and p = 0.19, which is a null result presented as a month of progress. An addendum published in April 2019 did not find an impact on Key Stage 2 results. Both findings belong in any honest account, and the second is almost never mentioned.

The compliance critique. Cushing (2021) studied how language policy operates in an English secondary school and argued that teachers are positioned as language policy managers working within a system of surveillance, compliance, coercion and control, explicitly naming the Teach Like a Champion industry. Applied to this routine, the charge is that a talk move with a cue word, a fixed posture, a timer and a required contribution can become a performance of participation rather than an occasion for thinking, and that what gets policed is compliance rather than reasoning. It is a serious argument and it is worth sitting with. Our view is that the risk is real and is a property of how the routine is used rather than of the routine itself: a prompt that invites genuine disagreement, and a teacher who changes the lesson based on what was said, is doing something the critique does not describe. A cue word attached to a question with one acceptable answer is doing exactly what the critique describes.

Talk can spread error efficiently. Covered above, and repeated here because it is the qualification most likely to matter in your classroom this week. Without a check, the routine distributes whatever the confident learner in each pair believes.

‍

Frequently Asked Questions

‍

References and Further Reading

Every study below was checked twice against the academic record: once to confirm the reference exists, and again to confirm the paper it resolves to is the paper being cited. Three candidate references were discarded during that second check because they resolved cleanly to entirely different studies. Each link goes to the original publisher rather than to an aggregator.

1. Wait-time and rewards as instructional variables, their influence on language, logic, and fate control: Part one, wait-time. Rowe, M. B. (1974). Journal of Research in Science Teaching, 11(2), 81-94. The origin of the three-second rule. Rowe found that extending the pause after a question changed the length, confidence and logical structure of what learners said. It is the cheapest change available to any teacher running paired talk.
2. The role of wait time in higher cognitive level learning. Tobin, K. (1987). Review of Educational Research, 57(1), 69-95. Reviews the wait-time research that followed Rowe and connects longer pauses specifically to higher-order thinking rather than to fluency alone.
3. Children's talk and the development of reasoning in the classroom. Mercer, N., Wegerif, R., and Dawes, L. (1999). British Educational Research Journal, 25(1), 95-111. Sets out disputational, cumulative and exploratory talk. The most useful single distinction for anyone designing a prompt, because it explains why agreeable pairs can talk for a minute and learn nothing.
4. Teaching children how to use language to solve maths problems. Mercer, N., and Sams, C. (2006). Language and Education, 20(6), 507-528. Children explicitly taught how to talk about problems got better at solving them. Children merely placed in pairs did not.
5. The teacher's role in promoting collaborative dialogue in the classroom. Webb, N. (2009). British Journal of Educational Psychology, 79(1), 1-28. What the teacher does during and immediately after paired talk turns out to matter more than how the pairs were formed.
6. Improving pupil group work interaction and dialogue in primary classrooms: results from a year-long intervention study. Baines, E., Rubie-Davies, C., and Blatchford, P. (2009). Cambridge Journal of Education, 39(1), 95-117. A year-long study in UK primary classrooms, and the strongest British classroom evidence in this list.
7. Classroom dialogue: a systematic review across four decades of research. Howe, C., and Abedin, M. (2013). Cambridge Journal of Education, 43(3), 325-356. Forty years of research reviewed in one place, including the field's own difficulty in showing that any particular arrangement of talk reliably raises attainment.
8. 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 by the Cambridge team, and the most useful paper here for a school deciding how much to invest in talk.
9. Language, discipline and 'teaching like a champion'. Cushing, I. (2021). British Educational Research Journal, 47(1), 23-41. Open access. The compliance critique, argued from fieldwork in an English secondary school. Read it before deciding what your talk routines are for.
10. The impact of nonverbal ability on prevalence and clinical presentation of language disorder. Norbury, C., Gooch, D., Wray, C., Baird, G., Charman, T., and Simonoff, E. (2016). Journal of Child Psychology and Psychiatry, 57(11), 1247-1257. The source of the two-learners-in-thirty figure, from a population screen of over 7,000 children.