Team Based Learning¶

Pattern · All patterns
Evidence · 25 claims (11 for, 12 mixed, 2 against) · 25 studies (8 causal, 7 quant-synthesis, 4 review, 2 associational, 2 design, 2 theoretical), q1–q4 · 8 of 25 report an effect size · 18 claims rest on one study

Description and scope¶

Team-Based Learning is a unit-level cycle in which learners study material before class, take a closed-book individual readiness test, retake the same test in a stable team with immediate feedback, appeal items in writing, receive clarification targeted at what the tests showed, and then spend most of class time on team application exercises. Its readiness-assurance sequence is a response-dependent policy: what the instructor does next depends on the individual and team responses. It instantiates part of the retrieval practice principle.

Which parts the retrieval evidence bears on, and which it does not. The individual test is a retrieval attempt after study, and the team retest supplies corrective feedback; the retrieval and feedback claims bear on those two steps, and on reusing readiness items later. They say nothing about team formation, team size or permanence, appeals, application exercises, simultaneous public reporting, peer evaluation or the grading weights. Those parts rest on other evidence (cooperative learning and active learning, below), on argument, or on nothing recorded in this wiki. No claim page in this wiki reports a study of Team-Based Learning itself. The study configurations below are evidence about components; the policy and its branches are an untested design proposal.

Inputs: establish the design brief¶

Field	Record or elicit
Learner and starting observation	Prior study of this content, the reading or media actually assigned, an individual response on a closed-book probe (item format, time since study), confidence, and language or access needs. Preserve unknowns; "did the reading" is an inference from responses, not an observation.
Context and activities	Class size, team size and how teams are formed, time per unit, item format (multiple choice is usual), how team feedback is given (scratch card or digital), what counts toward grades, and every earlier exposure to the items. Setting is usually `classroom`, duration `term-plus`.
Learner-valued goal	Ask what learners want from the unit and from working in a team, and why. Record disagreement with the designer objective; some learners value the grade, others a professional use.
Designer objective	The intended capability: recall of the pre-class content, its application to cases, teamwork, or several. Say which the readiness test is meant to assure, and which the application exercise is meant to build.
Outcome	Instrument, local criterion and horizon for each target: individual delayed recall of readiness content, individual performance on a new application item, a team product. A team score is not an individual outcome.

A request to "use TBL" does not specify these. Elicit them before choosing item counts, grading weights or a predicted gain.

Sequence and conditional policy¶

Set the course frame. Form teams and agree the grading scheme and peer evaluation. No retrieval evidence bears on this step.
Pre-class study. Learners study assigned material. Retrieval evidence assumes material has been studied; it does not say that reading alone is enough first instruction for every learner or topic.
Individual readiness test. Closed-book, individual, usually multiple choice. Retrieval evidence bears here: testing against restudy, item format, initial success. Multiple choice without feedback carries a recorded risk of learning the lures, which the next step is placed to correct.
Team readiness test with immediate feedback. The same items, retaken by the team until correct. Feedback evidence bears here, though team discussion and feedback arrive together and the studies separate neither from the other in this form.
Appeals. Teams argue in writing that an item was flawed or their answer defensible. No retrieval evidence bears on this step.
Targeted clarification. The instructor explains only what the individual and team responses show was not understood. Evidence on combining peer discussion with instructor explanation bears by analogy.
Application exercises, public reporting, peer evaluation. Teams work a shared problem and report simultaneously. Retrieval evidence does not bear on these. Active-learning evidence bears on lecture replacement in general, not on this exercise design.
Reobserve. Probe individual recall and application of earlier units' content at a delay, with new items. This step is a proposal added here; it is not part of the canonical sequence.

Observation	Discriminating probe	Proposed action, with its uncertainty
Individual score low, team score high on the same items	Was the material studied, or studied and not retained, or were wrong lures chosen? Give an individual short-answer probe on the missed items at the next session and ask what was read.	If the individual probe is now correct, the team feedback may have corrected it; keep the cycle and recheck later. If it is still wrong, the team result was carried by others: offer individual follow-up retrieval with feedback. Neither branch is validated.
Individual and team both wrong on an item	A shared misconception, a flawed item, or reading that did not cover it. Read the appeals, check the item against the reading, and ask a team to explain its choice.	Revise or drop a flawed item. Otherwise give targeted clarification on that concept and a new retrieval item on it later; do not reteach the whole reading.
Team reaches the correct option only after several scratches	Elimination by guessing, or discussion that changed members' understanding. Give an individual isomorphic item later with no discussion.	If the isomorphic item fails, treat the team result as not yet understood and plan another retrieval opportunity with feedback.
Readiness scores high, application exercise weak	Recall of facts without their use; an ambiguous problem; team process. Give an individual application item with the needed facts supplied.	If it succeeds with facts supplied, the gap is in use, not recall: work the application with modelling. If it fails, revisit the concept. Retrieval evidence does not predict this case.
One member dominates, others are silent	Team dynamics, uneven preparation or language demand. Compare members' individual test results and private explanations; read peer evaluations.	Adjust norms or roles, or team composition. This is a team-process judgement; the retrieval evidence offers nothing here.

Choosing configurations from evidence¶

Keep the individual test closed-book and low-stakes enough to be a learning event. Classroom quizzing [+M] improved classroom exam performance in eighth-grade science, college statistics and introductory psychology, including on completely new questions in one comparison of two sections; these are reported second-hand in one review chapter and not yet checked against their sources. None was a team-based course.
Follow multiple-choice readiness items with feedback. Multiple-choice tests without feedback [-M] led students to produce the lures they had selected on a later cued-recall or short-answer test, even with an overall benefit; second-hand from the same chapter, one of two entries passes the judge. This is the case for the team retest's immediate feedback, and for feedback on the individual test too if the team retest is ever dropped.
Expect feedback to matter most for errors. Feedback after retrieval [+M] raised retention after incorrect responses and added little after correct ones in a word-pair experiment, and classroom quizzes with feedback showed larger effects than those without, in a between-study comparison another meta-analysis did not reproduce. Team feedback in this pattern comes with discussion, so the studies do not isolate it.
Combine team discussion with targeted instructor explanation. In genetics courses using clicker questions, peer discussion [+W] improved performance on a later isomorphic question, and peer discussion followed by instructor explanation gave larger gains than either alone. These are classroom studies of peer instruction (one with no control group, one quasi-experimental), not of team readiness tests. One entry passes the judge on its abstract; the other could not be confirmed from its abstract.
Treat the application phase as active learning, not as retrieval. Active learning [~M] beat traditional lecturing on undergraduate STEM exams and concept inventories (0.47 SD across 158 studies), and only high-intensity active learning narrowed achievement gaps. Team-Based Learning is one form among many in those syntheses, which do not estimate it separately. Not settled on the abstracts.
Base team rewards on individual learning. Group rewards with individual accountability [+W] were associated with larger achievement effects (median +.32 against +.07) in a best-evidence synthesis of school studies, which also reports structured interaction working without rewards. These are school-age cooperative-learning studies, not this pattern; they bear on its grading, not its retrieval steps.

Do not rank these against one another, and do not read an effect from a laboratory word-pair or a lecture course as a forecast for a team-based course. A precise abstention names the missing comparison: no recorded study compares this sequence with the same content taught otherwise.

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/team-based-learning.json, scripts/synthesize_claim.py). The count is of distinct studies across the claim and the narrower claims that bear on it.

Does the team readiness test raise scores above the individual one? Students perform significantly higher on the team portion of the Readiness Assessment Test than on the individual portion in a TBL laboratory module [~M]: 3 distinct studies.
Which readiness score says something about later individual performance? IRAT performance shows a stronger association with final exam performance than TRAT performance does [~M]: 1 distinct studies.
High-stakes or low-stress readiness tests? Dental students in a revised low-stress TBL format scored higher on the individual summative assessment (iSAT) than medical students in the original high-stakes format, in a non-randomised comparison of two cohorts [~M]: 1 distinct studies.
How should team work be graded so every member is accountable? 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.
Permanent teams: who finds them hard, and how should teams be formed and run? 50% of neurodivergent students report some discomfort with semester-long teams compared to 22% of neurotypical students [~M]: 1 distinct studies.

Fitting the design to a situation¶

Ask for four facts when the brief omits them: how long the course runs (stable teams need weeks), what the readiness tests count for, class and team size, and whether any learners find permanent teamwork hard.

If the brief says…	Then change…	Basis
The instructor needs to know who prepared the reading	Use the individual readiness test, not the team test, as the signal of individual preparation; in one pharmacy course it tracked the final exam more closely	IRAT is more strongly associated with the final exam than TRAT [+W]; TRAT correlates weakly with lecture assessments [+W]
Team test scores are much higher than individual ones	Expect this; do not read the team score as each member's learning, and recheck individuals on the missed items later	Team RAT scores are higher than individual RAT [~W]
The cohort is in a high-stakes professional programme (medicine, dentistry)	Lower the grade weight of the readiness tests; in one non-randomised comparison a low-stress format gave similar readiness scores, less stress and higher individual summative scores	Low-stress format had higher iSAT scores [+W]; Low-stress format lowered stress and frustration [+W]
Readiness tests are multiple choice	Give immediate feedback on every team item (scratch cards or digital), and on individual items if the team test is dropped	Feedback enhances retrieval practice [+M]; Multiple-choice lures can be learned as false knowledge [-M]
Team work counts towards grades	Tie team credit to each member's individual performance, with peer evaluation, rather than to a single team product	Group rewards with individual accountability [+W]
Some learners are neurodivergent, or teams last a whole semester	Give teams defined roles, let students have a say in forming them, allow teams of three to five, and check in on team comfort during the term	Half of ND students report discomfort with semester-long teams [~W]; ND students prefer defined team roles [~W]
The team test shows a concept most teams missed	Give a short instructor explanation of that concept only, after the teams have discussed it	Peer discussion improves conceptual understanding [+W]
The course is a large lecture section	Keep the application exercises, which replace lecture time, but expect smaller gains than in classes of 50 or fewer	Active learning is strongest in small classes [~M]
The course is a laboratory with independent projects later	Use a TBL module on the core concepts before learners design their own projects	TBL module prepared students for independent projects [+W]
The brief is a one-off workshop, not a course	Run a single readiness cycle (individual test, team test with feedback, clarification, one application problem) with ad hoc teams; drop peer evaluation and appeals	untested proposal
The course runs live online	Run the individual test in the platform, the team test in breakout rooms with a shared answer form that gives immediate feedback, and simultaneous reporting by posting answers at once	untested proposal

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.

Assessment for learning improves achievement. [+S] — not settled: the text available could not confirm the entries (abstract)
Autonomy supports intrinsic motivation. [+M] — partly checked: 1 of 2 entries pass, the rest could not be confirmed (abstract)
Belonging interventions improve outcomes. [+M] — checked by the judge: all 2 entries pass (abstract)

Illustrative design instance and observation record¶

Invented for illustration; not a tested course. A second-year pharmacology unit: students read a chapter on one drug class, take a ten-item closed-book individual test, retake it as teams of six with scratch cards, and may appeal. The designer's objective is one-week recall of mechanisms and their use in choosing a drug for a case; several students say they value being able to explain a choice to a patient. The instructor sees most teams scratch twice on one interaction item while the individual test shows a common wrong option, clarifies that interaction for ten minutes, and runs a case exercise. A week later each student answers two new short-answer items on the unit, one recall and one application, with no notes.

Record: item and format → time since study, notes allowed → individual response and confidence → team response and number of attempts → appeal, if any → candidate interpretations → clarification given → individual response to a new item at the stated horizon. Keep individual and team records separate; a team score is not evidence about any one member. Preserve the learners' stated purpose and any change in it. An agent may summarise the record, but must not fill missing fields or assign probabilities to states without a calibrated model.

Elements and limits¶

Assigned readings, quizzes, retrieval practice, practice testing, feedback, peer discussion, argumentation, application, class discussion and peer assessment.

The retrieval evidence supports the readiness tests as retrieval with feedback; it does not support the claim that the whole cycle beats other designs, nor any team size, grading weight or application format. The response-dependent branches above, and the delayed reobservation step, are proposals to be tested with learners.

Claims¶

Cooperative Learning — shares positive interdependence and individual accountability; TBL adds the readiness assurance cycle and simultaneous reporting
Case-Based Learning — TBL's application phase is typically case-based; the Harvard method variant pairs naturally with TBL team structure
Cognitive Apprenticeship — complementary rather than competing: TBL assumes content is acquired before class, so expert modeling must happen in pre-class media or targeted mini-lectures

Examples¶

Medical education — pharmacology: Many medical schools (e.g., Vanderbilt, Duke-NUS) use TBL for pre-clerkship sciences: students complete assigned readings, take iRATs on drug mechanisms, then teams work patient-case application exercises choosing and defending treatment decisions.

Business education — management courses: Michaelsen's original context: teams of 5–7 in large organizational-behavior courses work through group-dynamics cases, with peer evaluation feeding into grades each unit.

Large undergraduate STEM courses: Introductory biology and chemistry courses at research universities have adopted TBL to convert lecture halls into team-application sessions, using digital tRAT tools such as InteDashboard or Top Hat for immediate team feedback. - Two-week fruitless gene TBL laboratory module on aggression and courtship in Drosophila melanogaster - Use IRAT as an indicator of subsequent assessment performance and exam-condition practice in pharmacy calculations

Key Sources¶

Michaelsen, L. K., & Sweet, M. (2008). The essential elements of team-based learning. New Directions for Teaching and Learning, 2008(116), 7–27. doi:10.1002/tl.330
Michaelsen, L. K., Knight, A. B., & Fink, L. D. (Eds.). (2002). Team-based learning: A transformative use of small groups in college teaching. Stylus Publishing. doi:10.5040/9798216023364
Parmelee, D., Michaelsen, L. K., Cook, S., & Hudes, P. (2012). Team-based learning: A practical guide. Medical Teacher, 34(5), e275–e287. doi:10.3109/0142159x.2012.651179
Haidet, P., Levine, R. E., Parmelee, D. X., Crow, S., Kennedy, F., Kelly, P. A., Brienza, R., Schneider, V., & Richards, B. F. (2012). Perspective: Guidelines for reporting team-based learning activities in the medical and health sciences education literature. Academic Medicine, 87(3), 292–299. doi:10.1097/acm.0b013e318244759e
Michaelsen, L. K. (2004). Getting started with team-based learning. In L. K. Michaelsen, A. B. Knight, & L. D. Fink (Eds.), Team-based learning: A transformative use of small groups in college teaching (pp. 27-50). Stylus Publishing.
Michaelsen, L. K., & Knight, A. B. (2004). Creating effective assignments. In L. K. Michaelsen, A. B. Knight, & L. D. Fink (Eds.), Team-based learning: A transformative use of small groups in college teaching. Stylus Publishing.
Demetry, C. (2010). Work in progress — An innovation merging "classroom flip" and team-based learning. 40th ASEE/IEEE Frontiers in Education Conference. doi:10.1109/FIE.2010.5673617

Page metadata

Field	Value
`type`	`pattern`
`id`	`team-based-learning`
`title`	`Team Based Learning`
`description`	`A unit-level readiness-assurance policy (individual closed-book test, team retest with immediate feedback, appeals, targeted clarification, then team application) whose test-and-feedback steps rest on retrieval evidence and whose team formation, application exercises and peer evaluation do not.`
`status`	`review`
`grain_size`	`unit, course`
`author`	`Larry K. Michaelsen`
`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.