Skip to content

Think-Pair-Share

Pattern · All patterns
Evidence · 17 claims (5 for, 12 mixed) · 35 studies (12 causal, 11 review, 6 quant-synthesis, 2 associational, 2 theoretical, 1 qualitative, 1 design), q1–q4 · 7 of 35 report an effect size · 7 claims rest on one study

Description and scope

A reusable policy for a short discussion routine on an open question: prompt → protected individual think time → partner exchange of reasoning → share outward to the larger group → synthesis. It instantiates the Peer Discussion principle, and it is the more general, lighter-weight routine beside the Peer Instruction pattern. Peer Instruction is a narrower configuration inside it: a conceptual question with one defensible answer, a committed individual vote, a revote and a new isomorphic question answered alone, which is the configuration the principle's evidence comes from. Think-Pair-Share is also used for interpretive or open prompts where no single answer and no isomorphic check exists; there the evidence below is thinner and the routine's outcomes must be specified locally.

No claim in this wiki tests Think-Pair-Share as a whole, or think time placed before pair discussion as a component. The study configurations below bear on its parts (pause length, peer discussion of conceptual questions, accountability in groups, discussion against lecture). The response-dependent policy is an untested design proposal. Use it to gather better observations of what learners think, not to certify that a class has understood because the share phase sounded good.

Inputs: establish the design brief

Field Record or elicit
Learner and starting observation What learners can say or write on this prompt before talking: a written note from the think phase is the cheapest observation. Prior instruction on the topic, language of discussion and anyone who rarely speaks. Preserve unknowns rather than labelling learners novice or "hesitant".
Context and activities Setting (classroom, online-instructor-led), class size, how pairs are formed, seconds or minutes given to each phase, whether the think phase is written, oral or silent, response mode for the share (oral, written, digital), and what the instructor does in synthesis.
Learner-valued goal Ask what learners want from the exchange and why. Record disagreement with the designer's objective; speaking up is not the same as agreeing with it.
Designer objective What the routine is for: a concept or principle to be revised, an interpretation to be argued from evidence, wider participation, or a check of understanding before moving on. These are different targets with different instruments.
Outcome The observation that will show it: an individual written answer and reason after the share, a new question answered alone, a rubric for reasoning quality, participation counts, and the horizon (immediate in class or delayed).

A request to "get more students talking" does not specify these inputs. Ask for them before prescribing phase lengths, a grouping rule or an expected gain. Check that the prompt supports more than a one-word answer; a recall question gives the think and pair phases nothing to work on.

Sequence and conditional policy

  1. Pose the prompt and protect the think phase. Give individual time before anyone speaks, and say how long. Ask for a written note or a committed answer where feasible: it is the only individual observation the routine produces before peers influence it. Record what was asked and how long the pause was.
  2. Pair to exchange reasoning. Ask each partner to give a reason and to say where the two disagree, not to agree on an answer. Record how pairs were formed. Circulate and sample what pairs say.
  3. Share outward. Decide in advance how pairs report: random selection of a reporter, written or digital submission, or volunteers. Volunteer-only sharing is an observation of who volunteers, not of the class.
  4. Synthesize. Connect what was shared to the objective, address the alternatives that came up and, where the prompt has a defensible answer, explain why the stronger reasoning is stronger.
  5. Reobserve individually. Where understanding is the aim, follow with a new question or a short written explanation answered alone, at the stated horizon. If observations disagree with the expectation, revise the prompt, timing or grouping rather than the interpretation.
Observation Discriminating probe Proposed action, with its uncertainty
Think-phase notes are blank or one word Ask a few learners privately what they took the prompt to ask; check whether the needed concept was taught. If the prompt was unclear or recall-only, revise it; if the concept is missing, explain or give an example before discussing. Longer think time alone may not help when there is nothing to think with; this is a proposal, not a tested branch.
Pairs report the same answer without reasons Ask a pair member to explain the partner's reasoning, or ask for a written reason from each learner. Add a reason requirement or sentence starters to the pair phase. Agreement in a pair may be copying, convergence or shared understanding; the share alone cannot say which.
The same few learners report in every share Compare who volunteers with think-phase notes from those who did not speak. Use random reporters or a written share mode. Silence may be uncertainty, a language or status barrier, or nothing to add; do not read it as disengagement without another observation.
One partner dominates the pair Ask each learner privately or in writing for their own reasoning after the pair phase. Vary pairing or assign roles (one explains, one questions, then swap). Talk volume does not establish learning for either partner.
Share sounds strong, individual follow-up is weak Compare individual answers on a new question with what the reporting pairs said. Treat the share as an observation of the reporters only; add instructor explanation and another individual check rather than declaring understanding.

Choosing configurations from evidence

  • For the think phase, longer teacher pauses [+W] are reported by two narrative reviews to be associated with longer, higher-cognitive-level student responses and changes in student and teacher attitudes, with no effect size given; both were read as abstracts and are not yet checked against their sources. The pause studied is the teacher's wait after a question and after a learner begins answering, not a structured individual think phase followed by pair talk, and the outcome is response quality in the moment, not later learning. It supports protecting think time, not a particular duration.
  • For the pair phase on conceptual questions, peer discussion on isomorphic clicker questions [+M] records undergraduate genetics students voting individually, discussing with neighbours, revoting and then answering an isomorphic question alone, with improved isomorphic-question performance even in groups where no one first answered correctly; a follow-up (n=150 majors and 62 nonmajors, quasi-experimental) found discussion alone and instructor explanation alone similar and the combination higher than either. No standardized effect size, immediate horizon, partly checked. It supports keeping both an individual commitment before pairing and a synthesis step after it; it says nothing about interpretive prompts without a defensible answer.
  • For what discussion may and may not do, discussion against lecture [~W] records that, in a classroom observation study, discussion-format students reported more time in active problem-solving and synthesis (8.3% against 1% of observed time) and less in passive and irrelevant thought, while reviews report lecture about as effective for transmitting information. Not yet checked against its sources. Expect the routine to change what learners do with ideas, not to beat explanation for delivering information.
  • For accountability, free-riding without individual accountability [~M] records that cooperative methods whose group rewards rested on each member's individual learning had a median effect of +.32 on achievement against +.07 for those without, in a best-evidence synthesis of 99 school studies lasting four weeks or more, while also noting that tightly structured peer interaction can work without rewards. Abstract-level for two of its three entries, not settled. These are multi-week cooperative programmes, not a two-minute pair exchange; transport only the point that each learner's own reasoning should be observable, as random reporters and individual follow-ups make it.
  • For pair size, pairs over triads [+W] is the authors' experience from teaching one astronomy inquiry twice, with no measured outcome. It is a reason to keep pairs, not evidence that pairs are better.

Do not convert a median effect from vote counts into a d, rank these by effect labels, or predict delayed retention from in-class share quality. A precise abstention names the missing comparison, target or observation and the next way to resolve it.

Fitting the design to a situation

Ask for three facts when the brief omits them: what kind of prompt it is (one defensible answer, or open), what the routine is for (revising a concept, wider participation, a check before moving on), and how the share will be collected in a class of this size.

If the brief says… Then change… Basis
The prompt is a conceptual question with one defensible answer and a known misconception Use the Peer Instruction configuration: a committed individual answer, pair discussion, a revote and a new question answered alone Peer discussion improves conceptual understanding [+M]
Many learners rarely speak, or are working in a second language Lengthen the think phase and ask for a written note before talking Longer wait time goes with higher-level responses [+W]
Learners have not yet been taught the needed concept (novice) Give a short explanation or example before the prompt; the think phase needs something to think with Minimal guidance is less effective for novices [~M]
One partner tends to do all the talking Assign turns: one explains, the other questions, then swap; ask both to say where they disagree Interactive engagement requires constructive turn taking [~W]
The brief asks whether to use pairs or larger groups Default to pairs, and offer trios where some learners prefer them; neurodivergent students in one survey were more comfortable in teams of three to five, neurotypical students in pairs Pairs worked better than triads in one inquiry [+W]; ND students prefer teams of 3–5 [~W]
Some learners find speaking to the whole class hard Let the share be written or reported by the pair, not a forced individual public answer ND students avoid courses requiring public presentations [~W]
The routine will be used weekly as graded group work Base any group credit on each learner's individual answer Group rewards with individual accountability [~M]
The prompt asks for interpretation of a text Teach a few discussion moves (asking for evidence, building on a partner's point) before using the routine; talk volume alone did not raise comprehension Some discussion approaches improve comprehension [~M]
The aim is only to deliver information Explain directly; use the routine where learners must do something with the information Discussion promotes more active thinking; lecture is no worse for information [~W]
The class is large and a whole-class share would take too long Call on randomly chosen pairs, or collect the share through a poll or written slips untested proposal
The class runs live online Use breakout pairs with a timed think phase in the chat or a shared document, and collect the share in writing untested proposal

The evidence by question

Further evidence, not yet read against this model

Claims this page cited before it was rewritten as a conditional model. They are evidence about the relationship, but none has yet been re-read against the model above; each says how far its own sources have been checked.

Illustrative design instance and observation record

Invented illustration, not a tested lesson. In an adult ethics class for trainee social workers, the instructor asks whether a worker should disclose a client's minor offence to a funder. Learners have two minutes to write a position and one reason; pairs then compare reasons and name the strongest objection to each; two pairs are chosen at random to report, and every learner submits a one-line revised position. One learner says she wants to be able to defend her view to a supervisor; the instructor's objective is reasoning that names the competing duties. Synthesis maps the reported reasons onto those duties, and a similar case is answered alone in writing the following week. These are local design choices; the evidence above does not establish their optimality.

Record: prompt and intended target → think-phase time and individual note → pair composition and sampled talk → share mode and who reported → synthesis given → individual follow-up answer → candidate interpretations (revised reasoning, adopted partner's view, unchanged) and their basis → next observation at the stated horizon. Preserve the learner's stated purpose and any change in it. An agent may summarize this record but must not fabricate missing inputs or assign numerical probabilities of understanding without a calibrated model.

Elements and limits

Individual reflection, discussion prompt, peer discussion, eliciting student thinking, sentence starters, class discussion and feedback. Principles the routine draws on: Peer Discussion, Formative Assessment, Purposeful Reflection and Immediate Feedback.

This pattern is scoped to a short routine embedded in a larger lesson. It is not a substitute for extended inquiry, seminar discussion or collaborative projects, and procedural fluency and recall need other configurations. The policy supports observation and design reasoning; the effect of the whole routine, of think time before pairing, of any phase length and of share mode on learning remains to be tested with learners.

Claims

Examples

  • A class concept question answered silently, discussed in pairs, and then synthesized whole-group.
  • Ethics or humanities prompts that move from individual stance to pair reasoning before broader discussion.
  • Quick checks in technical or professional training before the instructor debriefs the concept.

Key Sources

  • Lyman, F. (1981). The responsive classroom discussion: The inclusion of all students. In A. Anderson (Ed.), Mainstreaming digest (pp. 109-113). University of Maryland.
  • Kaddoura, M. (2013). Think pair share: A teaching learning strategy to enhance students' critical thinking. Educational Research Quarterly, 36(4), 3-24.
Page metadata
Field Value
type pattern
id think-pair-share
title Think-Pair-Share
description A reusable think, pair, share policy for open questions, in which protected individual think time and a partner exchange of reasoning are expected to widen and improve public responses, conditional on the prompt, pair dynamics, individual accountability and an immediate horizon.
status review
grain_size lesson
author Frank Lyman
canonical True
generated by claude/unspecified on 2026-10-02
trust unverified — no one has confirmed this page's content

Frontmatter is the page's machine-readable record — see the Schema & Guide.