Cooperative Learning¶

Principle · All principles
Evidence · 32 claims (13 for, 19 mixed) · 31 studies (12 quant-synthesis, 11 causal, 5 review, 2 design, 1 qualitative), q2–q4 · 12 of 31 report an effect size · 25 claims rest on one study

Conditional relationship¶

For learners of any age who can already take part in the target task, and for goals that benefit from explaining, checking and reconciling ideas (concepts, principles and complex skills more than verbal associations), structured small-group work, in which the group's success depends on every member's learning and each member's learning is assessed separately, is expected to raise individually measured achievement relative to individualistic or competitive arrangements over several weeks of use. Grouping alone is not this relationship: the expectation attaches to the structure (interdependence plus individual accountability), and science-classroom comparisons show it can fail. Evidence is mostly from school and undergraduate classrooms, on general or course achievement, at the end of multi-week programmes; it does not forecast a single session, an adult workplace team or a particular learner. The cooperative learning pattern specifies a reusable design policy; its local diagnostic branches remain proposals.

Fitting the design to a situation¶

Ask for three facts when the brief omits them: the learners' stage (primary, secondary, undergraduate, adult), how long the cooperative work runs (one session or four weeks or more), and how any group grade or recognition will be computed.

If the brief says…	Then change…	Basis
School-age learners, a unit of four weeks or more, an achievement goal	Recognise teams by the sum or average of each member's individual quiz results, announced in advance, rather than by a single group product	Group rewards on every member's individual learning [+M]
An undergraduate STEM course	Use structured small-group work in class sessions; expect a programme-level gain, smaller in studies where the investigator did not teach the course and not significant at two-year colleges	Small-group learning improves undergraduate STEM achievement [+M]
The plan is a jigsaw with a shared group grade	Add an individual quiz that counts toward the team's recognition; in secondary studies jigsaw was the least successful method, perhaps for lack of it	Secondary studies: STAD most, Jigsaw least successful [~W]
The goal is conceptual change in science	Do not assume groups beat pair or individual work; check each learner's concept individually, and coach student group leaders not to rush	Science-classroom studies are mixed [~W]; Poor group leaders undermined discussion [~W]
Groups work asynchronously online	Use roles to organise the group's process, but do not expect roles or rewards to raise achievement; check each member's posttest	Structuring interdependence did not change achievement online [~M]; Interdependence type shifted the kind of interaction [~W]
The group includes gifted or high-ability learners	Give them explaining roles and an individual challenge as well, and check their own gains; long-term evidence for this group is lacking	Cooperative learning and gifted students [~W]
The goal is reading comprehension in grades 3–4, with explicit strategy instruction available	Keep the explicit instruction central; the cooperative component did not add significantly over it with the same materials	The cooperative component's added benefit was not confirmed [~M]
A second-language class	Use cooperative groups for speaking practice, which gave learners more turns than teacher-led classes	Cooperative learning in language classes [+W]
Learners are primary pupils who need practice in a basic skill	Use structured peer tutoring with scripted roles rather than unstructured pairing	Peer tutoring improves achievement [+M]; Structured peer tutoring outperforms unstructured [+M]
Learners are trainee teachers and cooperation is itself an outcome	Run a structured cooperative technique and measure attitudes to cooperation separately from achievement; watch for unequal participation	A cooperative technique improved attitudes to cooperativeness [~W]; Unequal participation and time constraints noted [~W]
One session only	Use a brief structure with individual answers before and after the group step; do not expect the multi-week effect	untested proposal

Observation, state and explanation¶

Record what happens under stated conditions: the task, group size and composition, roles, how the reward or grade was computed, the talk the group produced, and each member's individual response afterwards. A group product is an observation about the group, not about any member's capability. "Contributes", "free-rides", "dominates" and "understands" are inferences from talk, work and individual checks; none is directly observed. A quiet member may be learning; a fluent one may be repeating a peer.

Response	Plausible explanations	Observation that could change the interpretation
Good group product, weak individual check for some members	Division of labour hid a member's gap; the member understood during talk but cannot yet perform alone; the individual check asks for something the group task did not	Compare each member's individual attempt on a task matched to the group task with one on a changed task; record who did which part of the product.
One member does most of the talking	Higher task-specific knowledge; higher status in the group; others are uncertain or excluded; the task is divisible so others are working silently	Ask each member privately to explain a step of the product; assign a rotating role and see whether participation and individual responses change.
Group talk is off task or socially focused	The task can be done by one person or needs no discussion; members lack the knowledge to discuss it; unclear purpose; ordinary relationship-building	Check whether an individual could complete the task alone; probe prerequisite knowledge individually; ask the group what they think the task is for.
Members reach a shared wrong answer	A shared misconception; deference to a confident peer; a reasonable reading of an ambiguous prompt	Ask members individually, before and after discussion, for their answer and reasoning; compare answers in a differently worded prompt.
Individual scores rise across a unit	Peer explanation; the individual quizzes and feedback themselves; extra time on task; the teacher's monitoring	Only a comparison that holds quizzes, time and teaching constant can attribute the gain to group structure. Record all of them.

These rows are proposals, untested with learners. A single probe shifts confidence among explanations; it does not identify a cause, and the probe itself is a learning exposure.

Evidence and qualifications¶

Structured cooperative goal structures against competitive and individualistic ones [+S]: a meta-analysis of 148 studies of early adolescents associates cooperative goal structures with higher achievement and better peer relationships (direction only; the abstract gave no magnitude), and an unpublished technical report on the Learning Together method reports weighted d = 0.91 against individualistic and d = 0.70 against competitive conditions. Checked by the judge against abstracts (all entries pass). It does not establish that any grouping works, gives no magnitude for cooperative learning in general, and does not separate the components.
Small structured groups in undergraduate STEM [+M]: 39 North American undergraduate studies, small-group against non-group instruction; achievement weighted d = 0.51, with persistence and attitudes also higher. Effects were larger when the investigator taught (0.73 against 0.41) and in two-group designs; two-year colleges showed a non-significant 0.21. Not settled against its source (abstract). It does not isolate interdependence or accountability: the included studies compared small-group with non-group instruction, and the restriction to structured groups is the claim page's own reading.
Group rewards tied to individual learning, against single-product or no group reward [~M]: a best-evidence review of 99 school studies of four weeks or more reports median effects of +.32 with group goals and individual accountability against +.07 without (vote counts and medians, not pooled estimates); a randomised 2×2 component study of 64 at-risk 4th and 5th graders found reward plus structured tutoring most accurate in computation. Not settled (0 of 3 entries confirmed from abstracts). The same review reports structured interaction that works without rewards, so it supports accountability as a moderator, not as strictly necessary, and it does not measure free-riding itself.
Collaborative writing [+M]: seven comparisons in grades 4–12 found jointly planned, drafted or revised texts of higher quality than texts written alone (weighted ES 0.75); one small randomised study of 10–11-year-olds found larger gains in individual writing after structured pair writing. Partly checked (1 of 2 entries pass, abstract). It shows the distinction between group product and individual capability matters; it is about writing, not cooperative learning generally.
Mixed science-classroom comparisons [~W]: a clearinghouse digest reports a grades 3–5 temperature study with no advantage over pair work, a preservice-teacher chemistry course where control classes won all 60 comparisons, and a seven-week biology study with no difference. Not yet checked against its source, second-hand, no effect sizes. It predicts that the expectation can fail in a given implementation; it does not say why.

Keep each study's learners, comparator, duration and outcome when transporting these observations. A pooled d is a programme-level average, not a learner's probability of benefit, and synthesis effects from unlike comparators should not be ranked.

The evidence by question¶

Each question below links a general claim that gathers the studies behind it into one table, one row per distinct study, with each study's design and effect as recorded and no averaged verdict (eval/synthesis/cooperative-learning.json, scripts/synthesize_claim.py). The count is of distinct studies across the claim and the narrower claims that bear on it.

Does structured cooperative group work raise achievement more than individualistic, competitive or traditional whole-class instruction? Cooperative Learning Improves Achievement [+S]: 10 distinct studies.
Does the effect differ by setting, outcome or learner? The effect of cooperative learning did not differ statistically by teaching style, outcome type, publication quality, bias status, or report type [~M]: 3 distinct studies.
Which structure matters: group rewards tied to each member's learning, assigned roles, or interdependence of any kind? In school studies, cooperative and peer-learning methods that reward groups on every member's individual learning raised achievement more than methods without such rewards [~S]: 7 distinct studies.
Does structured peer tutoring, a paired form of cooperative work, raise achievement, and does its structure matter? Peer Tutoring Improves Achievement [~S]: 4 distinct studies.

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.

Cognitive overload degrades learning when coordination demands exceed task benefits. [~M] — not settled: the text available could not confirm the entries (abstract)
Belonging interventions improve outcomes, particularly for marginalized students. [+M] — checked by the judge: all 2 entries pass (abstract)
Prospective teachers' qualitative feedback credited CCCT with communication, teamwork and engagement gains but noted unequal participation and time constraints [~W] — attached 2026-10-10 from Özgür Tutal (2026), which proposed "Structure cooperative learning so that group work embeds the five essential elements rather than merely placing students in groups".
Cooperative learning reduced emotional problems and increased peer relatedness for all students, not only marginalized students [+W] — attached 2026-10-23 from Van Ryzin et al. (2018), which proposed "Use cooperative learning as a sustainable, stand-alone component of bullying prevention that does not sacrifice instructional time".

Objective and learner-valued goal¶

Ask what each learner wants from the group work and why, separately from the designer's objective. A learner may value being heard, working with friends, finishing quickly or getting a good group grade; the designer may want every member to be able to explain the concept alone a week later. These can diverge sharply in group work, because a learner who values the group grade may rationally divide labour, which serves the product and undercuts the individual objective. Record agreement or divergence as an observation and negotiate the task, roles or grading when they conflict.

For example, a student in an undergraduate chemistry lab group may value a finished lab report and a fair share of the workload, while the designer wants each student to be able to predict and justify a reaction outcome on an individual test. A group report grade alone cannot show whether that objective was met; an individual prediction task, agreed in advance and scored on a stated rubric, can. The pattern shows how to keep both aims visible.

What would revise this model?¶

The expectation should weaken if comparable learners in structured groups, with both interdependence and individual accountability in place, fail to outperform matched individual or competitive conditions on individually measured, independent outcomes. If gains turn out to be carried by the individual quizzes, feedback or time on task rather than by group interaction, the model should credit those components instead. If structure without rewards works as well as rewards plus structure, as the review cited above suggests it sometimes does, drop rewards from the required core. Do not protect the model by relabelling every failure as "not real cooperative learning" after the fact; record the structure in advance.

Immediate group performance, individual achievement at the end of a unit, delayed retention, collaboration skill and belonging are separate claims. The present evidence does not establish an optimal group size, composition or reward scheme, effects for adult workplace teams or online groups, or benefit for every learner in a group.

Claims¶

Collaborative Learning — the broader category; cooperative learning is its most structured, accountability-driven form
Active Learning — cooperative structures are one of the most effective vehicles for active learning at scale
Assessment for Learning — individual accountability depends on frequent, low-stakes checks of each member's learning
Building Empathy — structured interdependence across diverse groups improves intergroup attitudes as well as achievement
Design laboratories to satisfy all five cooperative-learning conditions

Examples¶

Individual assessment during pair work: rotating roles, separate predictions and explanations, and independently-demonstrated ratings per outcome

Validated¶

Cooperative learning improves achievement when groups have positive interdependence and individual accountability. [+S] — Johnson & Johnson's (2009) synthesis of 60+ years of research found cooperative methods consistently outperformed competitive and individualistic conditions when both defining conditions were present.

Illustrative¶

Jigsaw — Each group member masters a unique segment of material in an "expert" group, then teaches it to their home group. Interdependence is structural: no member can complete the task alone. Aronson's original desegregation-era design remains one of the most replicated cooperative structures.

Think-Pair-Share — Learners think individually, rehearse with a partner, then share publicly. The individual think phase provides accountability; the pair phase provides elaboration. Low-cost and applicable to nearly any class discussion.

Peer Instruction (Eric Mazur, Harvard) — Students answer a concept question individually, discuss with neighbors to convince each other, then re-answer. ConcepTest discussion produces large conceptual gains in physics; the individual vote before and after discussion supplies accountability.

Complex Instruction (Stanford) — A status-treatment cooperative approach (Cohen & Lotan) using group-worthy tasks, assigned roles, and teacher interventions that publicly assign competence to low-status students, directly addressing free-riding and participation inequities.

PALS — Peer-Assisted Learning Strategies — A structured peer-tutoring program in reading and math in which dyads alternate tutor and tutee roles with scripted routines and point systems tied to individual improvement; validated in numerous randomized trials.

Key Sources¶

Johnson, D. W., & Johnson, R. T. (2009). An educational psychology success story: Social interdependence theory and cooperative learning. Educational Researcher, 38(5), 365–379. doi:10.3102/0013189X09339057
Springer, L., Stanne, M. E., & Donovan, S. S. (1999). Effects of small-group learning on undergraduates in science, mathematics, engineering, and technology: A meta-analysis. Review of Educational Research, 69(1), 21–51. doi:10.3102/00346543069001021
Slavin, R. E. (1996). Research on cooperative learning and achievement: What we know, what we need to know. Contemporary Educational Psychology, 21(1), 43–69. doi:10.1006/ceps.1996.0004
Johnson, D. W., Maruyama, G., Johnson, R., Nelson, D., & Skon, L. (1981). Effects of cooperative, competitive, and individualistic goal structures on achievement: A meta-analysis. Psychological Bulletin, 89(1), 47–62. doi:10.1037/0033-2909.89.1.47
Cohen, E. G., & Lotan, R. A. (2014). Designing groupwork: Strategies for the heterogeneous classroom (3rd ed.). Teachers College Press.

Page metadata

Field	Value
`type`	`principle`
`id`	`cooperative-learning`
`title`	`Cooperative Learning`
`description`	`For learners who can take part in the task, structured small-group work in which each member's learning is individually checked and counts toward the group's success is expected to raise individually measured achievement over several weeks, more than individualistic or competitive work; grouping alone is not expected to.`
`status`	`review`
`canonical`	`True`
`generated`	by `claude/unspecified` on `2026-10-01`
`trust`	unverified — no one has confirmed this page's content

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