Skip to content

Fostering Communities of Learning (FCL)

Pattern · All patterns
Evidence · 25 claims (14 for, 11 mixed) · 40 studies (13 quant-synthesis, 11 review, 6 causal, 5 qualitative, 3 theoretical, 2 associational), q1–q4 · 9 of 40 report an effect size · 14 claims rest on one study

Description and scope

A reusable policy for running a unit or course as a research-and-teach cycle: a shared inquiry theme → each learner joins a research group that becomes expert on one part of it → research groups re-form as teaching groups in which each member teaches their part (jigsaw) → reading and discussion run as reciprocal teaching, with roles handed from teacher to learners → a shared task that needs every part → individual checks and reflection decide the next cycle. The intended change is in each learner's conceptual-understanding of the whole theme, not only their own part, and, over several cycles, in their ability to question, explain and summarise without the teacher leading (self-regulation). A shared product, lively talk or a sense of community is not the target.

The pattern combines components that have their own pages and evidence: peer tutoring and learning by teaching [+M], reciprocal teaching, the jigsaw method and guided inquiry. It sits inside three converted relationships and does not restate them: the inquiry-based learning principle (guided inquiry, not unguided discovery, carries the benefit for novices), the collaborative learning pattern (joint work judged by what each member can do afterwards, with a move to cooperative learning's fixed interdependence and individual accountability when checks show free-riding), and the communities of practice principle, which names this page as its nearest pattern. That principle concerns adults joining a practice over a term or more; this page owns the classroom version, in which the teacher deliberately distributes expertise among learners who then teach it.

No claim in this wiki tests Fostering Communities of Learning as a whole, against ordinary instruction or anything else. The claims that bear on it test its components one at a time: preparing to teach and teaching, against studying for oneself; structured against unstructured peer tutoring; group rewards built on each member's individual learning; reciprocal teaching of reading; guided against unguided inquiry; and one secondary-school review in which jigsaw was the least successful cooperative method. Each is carried to this pattern by extrapolation, and the page says how far. The sequence below is the earlier page's five steps, made concrete and with the individual check given a step of its own, with each step's evidence status stated; the response-dependent branches are untested design proposals.

Inputs: establish the design brief

Field Record or elicit
Learner and starting observation Before the cycle, a short individual check on the whole theme (a concept map, three short explanations, or a quiz covering every subtopic), so later gains on parts a learner did not research can be seen. Record reading level and decoding fluency where texts carry the research, prior knowledge of each subtopic, and whether learners have taught peers before. Preserve unknowns rather than labelling learners novice.
The theme and its parts A theme that splits into 3–5 subtopics that are each learnable from available sources and that all matter for one shared task (for example, how animals in one habitat survive: food, shelter, defence, reproduction). If the parts do not depend on one another, learners have no reason to learn from each other.
Context and activities Setting (classroom, online-instructor-led, online-self-paced, workplace-clinical), time available (single-session, days-weeks, term-plus), number of learners and groups, how many adults are present to circulate, what sources learners can read, and whether all learners meet in one place.
Learner-valued goal Ask what learners want from the unit (understand the topic, do well on the test, make something others will use, work with friends). Record where it differs from the designer objective; the shared task can often carry both.
Designer objective Which target the assessment will weigh: understanding of the whole theme, depth on one's own part, the reading and questioning strategies themselves, or the shared product. Name one; the default is individual understanding of every part.
Outcome An individual test of every subtopic, scored by part (own part against parts taught by peers), immediate and at a delayed horizon; a separate measure of strategy use if reciprocal teaching is a goal; and, where it matters, a judged product. A standardised test and a test written for the unit can give very different pictures (see the reciprocal-teaching evidence below).

A request to "run the class as a learning community" does not specify these. Ask what theme, what parts, what individual test and how long, before forming groups.

Sequence and conditional policy

The default sequence is the earlier page's five steps, given a dose and a progression, with the individual check as a step of its own. Durations, group sizes and thresholds are proposals unless a claim is linked.

  1. Launch a shared theme and the individual check (1 session). Introduce the theme with a short teacher-led overview and a question the whole class must answer at the end; give the individual starting check (inputs, above). Evidence status: untested; from the earlier page. Guidance at the start follows the inquiry evidence [+M] that teacher-led inquiry in school science showed larger effects than student-led inquiry. Make the class question specific and give each learner an individual target on the check (for example, "explain all four survival strategies with an example of each"), not an open "explore the topic": specific, difficult goals [~M] outperformed vague "do your best" goals in Locke and Latham's narrative review, where learners were committed and able; it is goal-setting research, not a study of this pattern.
  2. Form research groups of 3–4, one subtopic each, with a role card and question stems (proposal). Each member reads and takes notes on the group's subtopic and is told from the start that they will teach it. Evidence status: preparing to teach and then teaching [+M] improved the tutor's own knowledge over studying with no expectation of teaching across 28 studies (g = 0.56; preparing alone g = 0.35), with interactive teaching better than non-interactive; read from abstracts, and the studies are not of this pattern.
  3. Research with reciprocal teaching (3–5 sessions of 30–50 minutes, proposal). Groups read their sources with the four reciprocal-teaching moves (predict, question, clarify, summarise). In the first cycle the teacher models each move on a common text and leads a group; by the second or third cycle a learner leads the discussion and the teacher circulates. Evidence status: reciprocal teaching [+M] improved reading comprehension in a review of 16 studies, with a median effect of .32 on standardised tests and .88 on tests the researchers wrote; the claim page's discussion says the routine needs sustained practice over several sessions and fluent enough decoding.
  4. Teach in jigsaw groups (1–2 sessions, proposal). Re-form groups so each holds one expert per subtopic; each expert teaches their part, with a structured protocol: explain, give an example, ask the group two questions, check understanding before moving on. Evidence status: structured peer tutoring [+M] produced larger achievement gains than unstructured tutoring (Cohen et al.'s 52 studies; structured roles also related to social and behavioural effects in 36 elementary studies). The protocol itself is untested. Limit: in one secondary-school review, jigsaw was the least successful cooperative method [~M], and its authors speculated this reflected the absence of group reward and individual accountability.
  5. Check every learner on every part, and make the group's recognition depend on it (proposal for the check; claim for the principle). A short individual quiz or explanation on all subtopics, scored by part; recognise teams on members' improvement over the starting check. Evidence status: group rewards built on each member's individual learning [+M] raised achievement more than methods without them in elementary and secondary school studies (median +.32 against +.07 in a review of 99 studies; d = 0.33 for peer-assisted learning overall), though the same review reports structured-interaction methods that work without rewards.
  6. Use the knowledge in a shared task, then reflect and set the next question (1–2 sessions, proposal). A task that needs every subtopic (design an animal for the habitat; write a case recommendation; build a class reference). Learners then compare their final check with their starting check and name what they still cannot explain, which seeds the next cycle's theme. Evidence status: untested; from the earlier page. Prompting learners to explain is supported in other settings: self-explanation prompts [~S] improved understanding on immediate tests by a moderate average amount (g = .55), with little evidence on delayed or classroom gains. Comparing the final with the starting check against a stated target is the monitoring step that self-regulated learning models [~M] treat as how learners notice gaps and change strategy (two syntheses, one theoretical; no effect size, and no test of it in a research-and-teach cycle).

Progression across cycles (proposal). A cycle runs 2–4 weeks; plan two or three per term. Cycle 1: the teacher chooses subtopics and sources, models every reciprocal-teaching move, supplies the teaching protocol and question stems, and leads synthesis. Cycle 2: learners lead reading discussions, the teacher circulates and steps in contingently; stems are offered, not required. Cycle 3: learners propose subtopics and sources from the questions left at the end of cycle 2, and lead synthesis with the teacher as one voice. A group moves to less support when most of its members score at least 70–80% on parts taught by peers and can lead a reading discussion without the teacher prompting each move (proposal). Raising and withdrawing support in this way follows contingent scaffolding [~M], whose evidence is mostly one-to-one tutoring; its claim page notes that genuine contingency across a whole class is hard to achieve.

Observation Discriminating probe Proposed action, with its uncertainty
Learners score well on their own part, poorly on parts taught by peers Ask a learner to re-explain a peer's part; ask the expert what they planned to say. If the expert's plan was thin, the teaching protocol is not carrying the content: add a check-for-understanding step and a one-page summary per part. If the plan was sound but the listener did not take it up, give listeners a question to ask and a note sheet. Neither branch is tested.
An expert teaches a misconception to the jigsaw group Before jigsaw, ask each research group to explain its part to the teacher in two minutes. If research groups converge on errors, add a teacher check before teaching (the learning-by-teaching claim page warns a tutor with gaps may rehearse misconceptions). Proposed.
Reading discussions stall or one member does all the moves Record who performs each of the four moves across a session. If moves are missing, the teacher models them again on the next text. If one member dominates, rotate the leader role by card each text. Untested.
Learners cannot read the sources Ask a learner to read a paragraph aloud and summarise it. If decoding is the problem, supply read-aloud or audio sources and simpler texts; reciprocal teaching's benefit is said on its claim page to depend on decoding fluency. Proposed.
The shared product is good, individual checks are weak Compare individual scores with the product grade, by member. Treat the product as the group's, not each member's; move to cooperative learning's individual accountability (team recognition from individual results) for the next cycle.
Some members contribute nothing in research or jigsaw, or are left to fetch materials while others explain Record, per learner, who explained, asked or summarised in one research and one jigsaw session; ask the quiet member to explain one point from their notes. If they can explain but are not heard, give them a bounded role (questioner, summariser) and make their part's explanation a required step before the group moves on; if they cannot, rehearse their part with the teacher first. Define in advance what a valued contribution from a quieter member looks like. Basis is thin: active participation mattering more than membership [+W] is an author's reading of a perception survey of 28 honours students, and the critique that nonparticipants are relegated to the fringes [~W] is a theoretical argument about communities of practice; neither tests a remedy.
The class reports enjoying the community but the delayed check shows no gain over the starting check Compare delayed scores by part and by learner. The community may be working as a social structure and not as instruction; shorten research, add direct teaching of the core ideas before research, and recheck. Proposed.

Fitting the design to a situation

The three facts that most change the decision:

  • Whether learners can learn a subtopic from the sources well enough to teach it. The pattern assumes each expert can research their part; unguided discovery of essential content learns less than explicit instruction, and a tutor with gaps may pass on errors. If unstated, ask: can these learners read the sources, and do they know enough of the domain to tell a good explanation from a wrong one?
  • Whether there is time for more than one cycle. The handover from teacher to learners, and the reading routine, need several sessions to take hold. If unstated, ask: how many weeks, and how many sessions a week?
  • What the assessment will weigh. Individual understanding of every part, the group's product, or strategy use each lead to different checks and recognition. If unstated, ask: what will each learner be tested on, and will anyone outside the class use the product?
If the brief says… Then change… Basis
Learners are novice in the domain Teach the core ideas directly before research; give each research group a short, pre-selected source pack and a note frame; check each group's understanding before jigsaw. Guided against unguided discovery [~S]: explicit instruction beat unassisted discovery, guided discovery beat other instruction; the meta-analysis does not break results down by prior knowledge, and is not of this pattern
Learners are younger readers, or struggling readers, in primary school Make reciprocal teaching the main activity of research, with the teacher leading one group per session and modelling each move for several sessions before handing over; keep subtopics to three. Reciprocal teaching [+M] (16 studies); especially poor readers [+W] (a narrative review citing one study). Neither tests the whole pattern
Learners are in upper secondary (grades 10–12) Do not rely on jigsaw alone: add team recognition built from individual quiz results, and a structured teaching protocol. Secondary review [~M]: only 33% of grade 10–12 studies were successful and jigsaw was least successful; group rewards on individual learning [+M]
postsecondary STEM course Use the cycle for one unit a term; groups of 3–4; individual checks each week; keep lectures for core ideas. Small-group learning in undergraduate STEM [+M] (d = 0.51 against non-group instruction; does not separate structured from unstructured groups)
Adults in a work setting (a professional learning cohort, workplace-clinical) Make the theme a recurring problem of practice; each member researches a case or source and teaches it back; the shared task is a protocol or guide the team will use; observe practice afterwards, not only self-report. Give the facilitator time to plan what each meeting asks members to reflect on, and treat disagreement about a case as material for the next round rather than a reason to close the meeting. Communities of practice principle, whose evidence is descriptive; mathematics PLC review [+M] (a review of descriptive studies, no comparison). One phenomenological study of three PLC leaders (Prytula 2012) reports that the PLC nurtured their metacognition [+W], that each leader's own metacognition shaped the work she led [+W], and one leader's view that disagreement catalysed learning [~W] (another adjourned to avoid conflict); a survey of small Chicago high schools found higher reported trust and collaboration [+W] (no effect size). None measures learning or practice against a comparison
single-session only Do not attempt the full pattern. Run one jigsaw on pre-read material with a teaching protocol and an individual check on every part; leave reciprocal-teaching handover out. Untested; reciprocal teaching's claim page says the routine needs sustained practice over multiple sessions
Mixed modes: some learners remote, some in the room Put each remote learner in a research group with at least one other remote learner or give the group a shared online document and audio channel; in jigsaw, the remote expert teaches from a shared screen with the protocol visible; collect individual checks online for everyone. Untested proposal
online-self-paced or phone-only, no live sessions Replace live jigsaw with asynchronous teaching: each expert posts a short explanation and two questions; peers answer and the expert replies; an automatic individual quiz on every part decides progress. Untested proposal; learning by teaching [~M] found interactive teaching better than non-interactive, so expect less benefit without live questions
One teacher for many groups (30+ learners) Use role cards and question stems so groups can run without the teacher; the teacher circulates on a fixed rota and checks each research group once before jigsaw; use written exit checks to decide which group to visit. Contingent scaffolding [~M]: evidence mostly from one-to-one tutoring; its claim page notes contingency across a whole class is hard. The rota is untested
Learners with low confidence or low literacy Give each learner a defined, bounded role (questioner, summariser) and a rehearsal with the research group before teaching; let experts teach in pairs; accept spoken or drawn explanations in checks. Untested proposal; positioning learners as sources [+M] widened participation in one lower-secondary science class (qualitative, no comparison group)
High stakes: an individual exam on all content Score every learner on every subtopic, add a teacher-led review of parts where peer-taught scores lag, and test at the delay that matters. Group rewards on individual learning [+M]; reciprocal teaching's gap between standardised and researcher-written tests [~M] is a reason to check with an independent measure
A product or performance for a real audience (an exhibition, a guide for younger pupils, a report for a partner) Make it the shared task in step 6 and name the audience at launch; still score individual understanding separately from the product. Untested proposal; extends the earlier page's step of creating artifacts and sharing findings

Nine rows cite a claim, each carried beyond the population or design its claim page reports, as its Basis cell says; two of those (self-paced online and low confidence or literacy) are otherwise untested proposals. Three rows (a single session, mixed modes, and a real audience) are untested.

Choosing configurations from evidence

  • The teaching role. Learning by teaching improves tutor learning [+M]: across 28 studies, preparing to teach and then teaching improved the tutor's domain knowledge over studying with no teaching expectancy (g = 0.56), preparing alone less (g = 0.35), with interactive teaching better and gains persisting after a delay; a meta-analysis of 65 school tutoring programmes found tutors also gained understanding of the subject. Both read from abstracts. Teaching others as a generative strategy [+M] rests on the same two studies and adds none. They support assigning each learner a part to teach and making teaching interactive; they do not show that learners learn the parts taught to them by peers as well as they would from a teacher.
  • Structure in peer teaching. Structured peer tutoring outperforms unstructured [+M]: structured programmes produced larger achievement gains in Cohen et al.'s synthesis (overall d = .40; the moderator's size not reported), and structured roles related to social and behavioural effects in 36 elementary studies (abstract only). Peer tutoring improves achievement [+M] adds a mathematics meta-analysis of 50 studies (g = 0.333). Carry these to the jigsaw teaching step as "give experts a protocol and defined roles", no further.
  • Accountability. Group rewards built on each member's individual learning [+M]: in school studies, cooperative methods rewarding groups on every member's individual learning raised achievement more than those without (median +.32 against +.07; 99 studies), with peer-assisted learning effects larger under interdependent rewards (d = 0.33 overall). Slavin's review also reports structured-interaction methods working without rewards, so rewards are one route to accountability, not a requirement.
  • Reciprocal teaching. Reciprocal teaching improves reading comprehension [+M]: median .32 on standardised tests against .88 on researcher-written tests across 16 studies; one review, read as an abstract. The benefit's size depends heavily on the measure. The narrower claim for poor readers [+W] is a narrative review's report of one study, with no effect size.
  • Guidance in the research phase. Teacher-guided inquiry outperforms student-led [+M]: in a meta-analysis of 37 school science studies, teacher-led inquiry showed effects about .40 larger than student-led; a synthesis of 138 K–12 studies found a positive trend for inquiry over passive instruction for conceptual understanding. Guided against unguided discovery [~S] marks the boundary: unassisted discovery learned less than explicit instruction (d = 0.38 favouring explicit instruction), and discovery with feedback, worked examples or scaffolding did better than other instruction. A research group left with sources and no support is the unguided case.
  • The limiting finding. In secondary schools, jigsaw was the least successful cooperative method [~M]: a review of 27 secondary studies found cooperative learning favoured in 68% of 37 comparisons, STAD the most consistently successful and jigsaw the least, and only 33% of grade 10–12 studies successful. The authors' explanation (no group reward or individual accountability) is a speculation, not a tested moderator. This is the component of this pattern with the weakest record here, which is why the sequence pairs jigsaw with individual checks and a teaching protocol.

Do not rank Fostering Communities of Learning against direct instruction, cooperative learning or ordinary inquiry from these claims: none compares them, none tests the components together, and none measures the time a research cycle costs. Do not read a confident class discussion, a good group product or learners' sense of belonging as each learner's understanding. A precise abstention names the missing comparison (the research-and-teach cycle against the same content taught directly with individual practice, on individual understanding of every part at a stated delay) and the local observation that would stand in for it: scores on parts learners did not research.

The evidence by question

Further evidence, not yet read against this model

Read against the model on 2026-10-19. Of the eleven claims listed here then, ten are now used above: contingent scaffolding and self-explanation (already in the sequence and the progression), self-monitoring and specific goals (steps 6 and 1), active participation and the critique of nonparticipation (the observation table's row on members who contribute nothing), and the four professional-learning-community claims (the adults-at-work row). What stays here does not change the model.

Illustrative design instance and observation record

Invented illustration, not a tested unit. A Year 5 class of 28 (primary), one teacher, one teaching assistant, four weeks on "how animals survive in the rainforest", assessed by an individual end-of-unit test and a class guide for the Year 3 class. Several pupils read below age level, so the table changed the default: the teacher spends the first session on a short direct lesson and the starting check; research groups of four take food, shelter, defence or reproduction, each with a short source pack and audio versions; for three sessions the teacher leads one group a session in reciprocal teaching while the assistant circulates with a rota; before jigsaw each group explains its part to the teacher in two minutes, and the defence group's error about camouflage is corrected; jigsaw groups teach with a protocol card (explain, example, two questions, check); a ten-question quiz covers all four parts and teams are recognised on improvement over the starting check; the shared task is the Year 3 guide. One pupil says she wants to "draw the animals", and her group's guide uses her drawings for every part; the designer's objective is understanding of every part. The scores show pupils weakest on parts taught by peers, so the second cycle adds a listener note sheet. None of the linked claims tested this configuration, population or audience task; these are local choices to observe.

Record: theme and parts → starting check by part → research group, sources and support given (who modelled which move, for how many sessions) → expert's pre-jigsaw explanation and corrections → jigsaw teaching (protocol used, questions asked) → individual check by part (own part, peer-taught parts) → recognition rule → shared task and its audience → candidate interpretations (expert's explanation thin, listener not taking up, sources unreadable, accountability missing) and their basis → reobservation at the stated delay. Preserve each learner's stated purpose and any change in it. An agent may summarise this record but must not fabricate missing inputs or assign numerical probabilities of understanding without a calibrated model.

Elements and limits

Peer collaboration, reciprocal teaching, peer discussion, knowledge-building discourse, reflection and formative assessment; the jigsaw classroom strategy. Principles the pattern draws on: Communities of Practice, Inquiry-based Learning, Peer Discussion, Mentoring/Coaching, Scaffolding and Fading and Social Interdependence.

This pattern is scoped to a class or cohort that researches the parts of a shared theme and teaches them to one another, over several weeks. Short coverage of a fixed syllabus, content learners cannot learn from available sources, and goals that are one learner's skill rather than shared understanding need other configurations (direct instruction, cooperative learning, peer instruction). Whether the full research-and-teach cycle produces better individual understanding than its components alone, or than direct teaching of the same content, for which learners, at what time cost and at what horizon, remains to be tested.

Claims

Examples

  • Science communities where learners investigate subtopics and teach findings back to the class.
  • Humanities courses where groups develop domain expertise and contribute to shared interpretive discussions.
  • Professional learning cohorts that collectively build a shared evidence base around practice problems.
  • History learners examining different primary sources and teaching their significance to peers.
  • Science groups becoming experts on different parts of a system before building a shared explanation.
  • Professional learning cohorts dividing complex policy or process documents for peer teaching.

Key Sources

  • Brown, A. L., & Campione, J. C. (1996). Psychological theory and the design of innovative learning environments: On procedures, principles, and systems. In L. Schauble & R. Glaser (Eds.), Innovations in learning (pp. 289-325). Erlbaum.
  • Scardamalia, M., & Bereiter, C. (2006). Knowledge building: Theory, pedagogy, and technology. In R. K. Sawyer (Ed.), The Cambridge handbook of the learning sciences (pp. 97-115). Cambridge University Press. doi:10.1017/cbo9780511816833.008
  • Aronson, E., & Patnoe, S. (2011). Cooperation in the classroom: The jigsaw method (3rd ed.). Pinter & Martin.
  • Aronson, E., Wilson, T., & Akert, A. M. (2001). Social psychology (4th ed.). Prentice Hall.
  • Slavin, R. E. (1994). Cooperative learning: Theory, research, and practice (2nd ed.). Allyn & Bacon.
  • Hattie, J. (2009). Visible learning: A synthesis of over 800 meta-analyses relating to achievement. Routledge. doi:10.4324/9780203887332
  • Arduini-Van Hoose, N. (2020). Cooperative learning. In Educational psychology. Retrieved from https://edpsych.pressbooks.sunycreate.cloud. CC BY-NC-SA 4.0.
  • Aronson, E., & Patnoe, S. (2011). Cooperation in the classroom: The jigsaw method.
  • Aronson, E., & Patnoe, S. (2011). Cooperation in the classroom (3rd ed.). Pinter & Martin.
  • Prytula, M. P. (2012). Teacher Metacognition within the Professional Learning Community. International Education Studies. https://doi.org/10.5539/ies.v5n4p112
  • Stevens, W. D., with Kahne, J. (2006). Professional Communities and Instructional Improvement Practices: A study of small high schools in Chicago. Consortium on Chicago School Research at the University of Chicago. https://consortium.uchicago.edu/publications/professional-communities-and-instructional-improvement-practices-study-small-high
  • Locke, E. A., & Latham, G. P. (2002). Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. American Psychologist, 57(9), 705–717. doi:10.1037/0003-066X.57.9.705
  • Butler, D. L., & Winne, P. H. (1995). Feedback and self-regulated learning: A theoretical synthesis. Review of Educational Research, 65(3), 245-281. doi:10.3102/00346543065003245
Page metadata
Field Value
type pattern
id fostering-communities-of-learning-fcl
aliases jigsaw, jigsaw-learning, jigsaw-method, plc-facets-catalyze-scholarship-epistemology
title Fostering Communities of Learning (FCL)
description A reusable research-and-teach cycle in which learners become expert on one part of a shared theme, teach it to peers in jigsaw groups and read through reciprocal teaching, expected to raise each learner's understanding of the whole theme where research is guided, teaching is structured and every member is checked on every part; no claim tests the pattern as a whole, only its components.
status review
grain_size course
author Ann L. Brown and Joseph C. Campione
canonical True
generated by claude/unspecified on 2026-10-19
trust unverified — no one has confirmed this page's content
sources 1 — prytula-2012

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