Skip to content

Structured Academic Controversy

Pattern · All patterns
Evidence · 10 claims (10 mixed) · 24 studies (8 quant-synthesis, 6 causal, 6 review, 2 theoretical, 1 associational, 1 qualitative), q1–q4 · 6 of 24 report an effect size · 2 claims rest on one study

Description and scope

A reusable policy for cooperative argument on a contested question: a question that admits more than one defensible answer → each pair in a small group prepares one assigned position from evidence → each pair presents and the other pair listens and restates → open discussion → the pairs reverse positions and argue the other side → the group drops advocacy and writes a joint, reasoned conclusion → each learner states and justifies their own position alone. It is the cooperative configuration of argument. It belongs to the family the cooperative learning pattern describes, under its principle: a group whose shared product depends on each member's work, with each member's learning checked separately. It sits beside Debate, which keeps sides fixed and ends with a judged winner. SAC adds two things debate lacks, the role reversal and the joint synthesis, and the target changes with them. Debate targets a case defended under challenge. SAC targets an integrated position, together with the ability to state the opposing case fairly. The cooperative learning model's rule (check members individually; do not read the group product as anyone's capability) applies here unchanged. This page adds the argument-specific policy around it: the controversy, the reversal and the synthesis.

No claim in this wiki tests Structured Academic Controversy. It is not tested against debate, against discussion, or against no activity, and no claim tests the reversal or the synthesis step on its own. The claims that bear on it are of three kinds. Syntheses of cooperative goal structures and of group rewards are about cooperative learning in general. Argumentation and discussion studies are not SAC. One review names "consider the opposite" as a debiasing strategy, and the reversal step resembles it. The page carries these over to SAC by extrapolation and says how far each may be carried. The sequence and response-dependent policy below are an untested design proposal. Use them to gather better observations of how each learner reasons about a contested question, not to certify understanding because the group reached consensus.

Inputs: establish the design brief

Field Record or elicit
Learner and starting observation A short individual written position on the question, with the best reason for it and the best reason against it, collected before positions are assigned. Prior knowledge of the topic (can they name evidence on both sides unaided?), language of instruction, experience of public argument, and anyone for whom the topic is personal. Preserve unknowns rather than labelling learners novice or "confident".
The question Whether it supports two or more evidence-bearing positions that informed people hold, or would create a false equivalence (one side has no credible evidence) or set learners against a group present in the room.
Context and activities Setting (classroom, online-instructor-led, online-self-paced), single-session or days-weeks, group size and how pairs are formed, how positions are assigned, the source pack or research time, turn lengths, whether a facilitator circulates, and how the joint product and any grade are computed.
Learner-valued goal Ask what learners want from it: to work out what they think, to defend a view they already hold, to understand people who disagree, a good grade, or getting it over with. Record disagreement with the designer objective.
Designer objective Which capability is the target: stating the opposing case fairly (perspective-taking), building claims from evidence, weighing evidence into a reasoned judgment, knowledge of the topic (concept, principle), or a disposition (attitude-motivation). Name the one the assessment will weigh.
Outcome Instrument and horizon: an individual written argument after the session on the same question or a new one (near-transfer, far-transfer); a fair restatement of the opposing case scored on a rubric; conceptual-understanding of the topic; or a self-report of attitude (motivation-affect). Say which is immediate and which delayed. The joint report is not this outcome.

A request for "a balanced discussion of a controversial topic" does not specify these inputs. Ask for the target, the instrument and the question before choosing group size, timing or grading.

Sequence and conditional policy

This is the earlier page's default sequence, kept because no claim tests a SAC sequence. The quantities are proposals for one 60–90-minute session and are labelled as such. Each step says what bears on it.

  1. Choose and check the question. It must admit at least two defensible, evidence-bearing positions. If one side has no credible evidence, do not run SAC: teach the evidence, or run a discussion of why the question is settled. Evidence status: untested; from the earlier page (requirement "a genuinely controversial question", constraint "false equivalence risk").
  2. Collect each learner's starting position in writing: one position, one reason for, one against (about 3 minutes). Evidence status: untested proposal. Its job is to give the synthesis and the individual check something to compare with.
  3. Form groups of four (two pairs), assign each pair a position, and give a source pack covering both sides. Allow about 10–15 minutes of preparation in class, or a prior homework reading. Groups of four are an untested proposal: the earlier page says "pairs or teams". If the class does not divide into fours, use a three (one learner observes and rotates in at the reversal), not a single large group. Evidence status: untested; from the earlier page. Supply argument frames or sentence starters where learners are young or new to the format. Argumentation improves reasoning [~M]: the claim page's discussion says scaffolded argument is where benefits are reported, but that is the page's reading, not its studies'.
  4. Present, listen and restate (about 4 minutes per pair). Before responding, the listening pair restates the other pair's best argument until the presenters accept it. Evidence status: untested; from the earlier page ("present and challenge positions", "restate the other side fairly").
  5. Open discussion (about 5–8 minutes): challenge the evidence, not the people. Evidence status: untested. The facilitator watches whether talk engages the other case. Structured discussion methods [~M] report that more talk alone did not drive gains, so the length of the discussion is not the thing to maximise.
  6. Reverse positions (about 4 minutes per pair). Each pair argues the side it has just heard, adding at least one point the other pair did not make. Evidence status: untested; from the earlier page ("role reversal is central"). Considering the opposite [~W] is the nearest claim. It is about judgment bias under a prompt, not about arguing an assigned side.
  7. Drop advocacy and write a joint conclusion (about 10–15 minutes). It must state the points of agreement, the tension that remains unresolved, and the group's best-supported judgment with its evidence. Consensus is not required; a reasoned "we disagree because…" is an acceptable product. Evidence status: untested; from the earlier page ("weak synthesis undermines the pattern").
  8. Check each learner individually (about 5 minutes, or as homework). Ask for a revised position, the strongest point against it, and what changed. If the target is transfer, ask for an individual argument on a new question at the stated horizon. Evidence status: individually checked contributions [~M] bear on this step, from school studies of four weeks or more. They are not SAC.
Observation Discriminating probe Proposed action, with its uncertainty
A pair cannot restate the other side's argument, or restates a weak version of it Ask one member, away from the group, for the other side's strongest piece of evidence; compare with the source pack. If they cannot find it in the sources, the gap is knowledge: give the evidence and a listening template before the reversal. If they can find it but did not say it, the problem is advocacy: require an accepted restatement before discussion continues. Neither branch is tested.
Role reversal produces a caricature (a weak, mocking or parroted version) Ask for one new point for the reversed side that the original pair did not make. If they can produce one, the caricature may be reluctance; name it and continue. If they cannot, give preparation time with the sources for the reversed side. Do not count a reversal performed as evidence of perspective-taking.
The joint conclusion is a vote, a split-the-difference compromise, or "both sides have a point" Ask the group which piece of evidence their conclusion rests on and what would change it. If no evidence is named, return the group to the synthesis with a frame (agree / disagree and why / best judgment and its evidence). A shallow synthesis is the earlier page's named failure; it is not measured here.
Group consensus, but individual positions afterwards are unchanged and unjustified, or simply repeat the group text Compare each learner's starting and revised positions and reasons; ask each to defend one sentence of the joint report alone. If learners cannot defend the group's conclusion, treat the report as the group's and give individual follow-up work. Consensus is not individual learning.
Distress, personal attack, or a learner for whom the question is about their own identity Ask privately; check whether the question is a genuine academic controversy or one that sets a group in the room up as one "side". Reframe the question around policies or interpretations, not people; allow a written or observer role; or replace SAC with a different format for this topic. The earlier page names this risk; no claim here measures it.

Fitting the design to a situation

The three facts that most change the decision:

  1. Whether learners know enough about the topic to argue both sides from evidence. Without that, preparation becomes research and the reversal becomes a guess. If the brief does not say, ask: "Could they name one piece of evidence on each side today, unaided?"
  2. What the assessment will weigh: an individual's reasoned position, a fair account of the other side, topic knowledge, or the group's product. That decides whether steps 2 and 8 are optional or the core. If the brief does not say, ask: "What will each learner have to produce alone afterwards, and when?"
  3. Whether the question is personal or charged for anyone present. That decides whether assigned sides and a reversal are safe at all. If the brief does not say, ask: "Is anyone in the group likely to be argued about, rather than argue?"
If the brief says… Then change… Basis
Expertise novice on the topic (cannot name evidence on either side) Lengthen preparation (a prior reading, or 20+ minutes with a structured source pack per side) and give an argument frame. Shorten open discussion. Defer SAC until a short explicit input on the evidence has been given. untested proposal; argumentation improves reasoning [~M] names prior knowledge as an open question on its claim page, not a tested moderator
Expertise advanced (graduate students, professionals who know the debate) Cut preparation and the source pack. Spend the time on the reversal and synthesis. Raise the synthesis bar: name the evidence that would settle the disagreement. untested proposal
Age child (upper primary) Use concrete, local questions (a class or school rule), pairs of two not four, sentence starters for restating ("You think… because…"), an oral synthesis recorded by the teacher, and shorter turns (about 2 minutes). untested; from the earlier page ("topics can be chosen to match learner maturity", "sentence stems"); argumentation [~M] reports larger effects in K-12 than higher education for online science argumentation, untested for SAC with children
Age adolescent or adult in civics, ethics, history or policy Use the default sequence. For adults, ask for the learner-valued goal explicitly, since adults often arrive with a held view; collect the starting position privately. untested; from the earlier page (target learners); untested proposal for the adult change
Goal is concept or principle knowledge of the topic, not argument Put the content check in the source pack and in step 8 (an individual item on the evidence itself). Do not assume discussion delivered the content. structured discussion methods [~M]: only a minority of discussion approaches improved comprehension or critical thinking (general discussion, not SAC)
Goal is attitude-motivation (tolerance, openness to the other side) Measure it separately, with a self-report or behaviour, at a delayed horizon. Do not infer it from a fair restatement in the session. Expect the effect of a single session to be small or short-lived. untested proposal; perspective-taking and attitudes are not tested for SAC in this wiki
Setting online-instructor-led (synchronous video) Use breakout rooms of four with a shared document whose columns are position / restatement / reversal / joint conclusion. Visit each room at the reversal. Give the timing in the document. untested proposal; argumentation [~M] includes online science argumentation, not SAC
Setting online-self-paced, or no teacher present Run it asynchronously over days-weeks: positions posted, restatements as replies that the original author must accept, reversal posts, then a joint document. Grade the restatement and the individual position, not post counts. A shared misconception and a caricatured reversal will go unseen, so add a facilitator check before the synthesis. untested proposal
single-session only (under 45 minutes) Keep restate, reverse, synthesise and the individual check. Cut open discussion and do preparation from a one-page source sheet. Do not cut the synthesis or the individual check: they are what distinguish SAC from debate. untested; from the earlier page ("strongest when the synthesis stage is treated as the instructional endpoint")
High stakes: individual grades, or a joint report graded for the group Grade each learner's own position and restatement. If the joint report counts, make it a small share, or score it on each member's ability to defend a part chosen at random. individual accountability [~M], school studies of four weeks or more; untested for SAC
The topic is charged for someone present, or learners argue in a second language Choose a question about a policy or interpretation, not a group of people. Let learners pick which position they prepare first. For learners working in a second language, give bilingual or simplified sources and allow written restatement. untested; from the earlier page ("emotional intensity", "performance pressure"); untested proposal for the language change

Four rows cite a claim in their basis (child, concept goal, synchronous online, high stakes), each carried by extrapolation from argumentation, discussion or cooperative-learning evidence; none tests SAC. The novice row cites a claim only to say the moderator is untested, and the other rows are untested proposals or come from the earlier page.

Choosing configurations from evidence

  • Cooperative goal structure, the family SAC belongs to. Cooperative learning improves achievement [~M]: a meta-analysis of 148 studies of early adolescents associates cooperative goal structures with higher achievement and better peer relationships than competitive or individualistic ones (direction only; the abstract gave no magnitude). An unpublished technical report gives weighted d = 0.91 against individualistic and d = 0.70 against competitive conditions for one method, Learning Together. It is checked by the judge against abstracts. It does not test SAC, the reversal or the synthesis. Carry it as far as "SAC is a cooperative structure", and do not use it as an estimate for SAC.
  • Accountability, not the joint product. Group rewards tied to individual learning [~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). The same review reports structured interaction working without rewards. Not settled against its sources (abstracts). It supports step 8 and the grading row. It does not say a single SAC session gains anything.
  • Structured argument with peers. Argumentation improves reasoning [~M]: a meta-analysis of 46 online science argumentation studies (5,415 students) reports g = 0.603 on argumentation skill and g = 0.596 on content learning, larger in K-12 than higher education; it was read from the abstract, so the comparison conditions are not established. A multiyear dialogic-argumentation intervention with young adolescents, not described as randomised, transferred to better individual essays on new topics than a comparison group given essay practice and whole-class discussion; no effect size is printed. Neither is SAC, and neither isolates reversal or synthesis. It supports an individual written argument as the outcome to check.
  • Discussion is not the active ingredient. Structured discussion methods improve comprehension [~M]: its claim text names structured academic controversy, but neither entry tests it. A meta-analysis of classroom discussion approaches found only a minority improved comprehension or critical thinking, moderated by design, measure and student ability (abstract only, no pooled effect read). A year-long single-group Quality Talk study of 35 fourth-graders had no control group. Its merge candidate, discussion quality drives comprehension [~M], carries the same two entries. It limits any SAC design that counts talk or participation as the outcome.
  • Consider the opposite. Considering the opposite reduces bias [~W]: a systematic review names the strategy as previously validated in its background, but its own data do not isolate it, and no effect size is reported. It motivates step 6 and the "strongest point against" prompt in step 8. It is about judgment bias, not about argument skill or attitudes toward the other side.

Do not rank SAC against debate, Socratic seminar or discussion from these claims: none compares them. Do not read consensus, a lively discussion or a fair-sounding reversal as a learning gain. Do not predict delayed attitude change from a single session. A precise abstention names the missing comparison: SAC against debate or plain discussion, on an individually measured argument or perspective-taking outcome, at a delayed horizon. It also says what local observation would stand in for it (each learner's starting and revised position, scored on one rubric).

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. The evidence above comes mostly from school classrooms and undergraduate courses. This instance is a workplace one: twelve hospital ward nurses (adult, workplace-clinical, intermediate on the topic) in a 60-minute in-service session. The question is whether the ward should move from fixed to self-rostered shifts, and the hospital is consulting on it. The table changed four things. The adult row meant each nurse's starting position was collected privately, by a two-line anonymous note, because most already hold a view and some are union representatives. The charged-topic row meant the question was framed around the policy and its evidence (a two-page summary of published rostering evaluations and the ward's own sickness-absence figures), not around colleagues' preferences, and anyone could take an observer role. The single-session row cut open discussion to 5 minutes and kept the reversal and the synthesis. The stakes row meant nothing was graded. The joint conclusion went to the ward manager as the group's view, labelled as such, and each nurse wrote their own revised position privately. One nurse said she wanted to understand why night staff opposed the change; the educator's objective was weighing evidence rather than anecdote. Both were recorded.

Record: question and intended target → each learner's starting position and reasons → pairing and assignment → sources given and preparation time → restatement accepted or not, and after how many attempts → reversal: new points added or not → joint conclusion: evidence named, tension stated → each learner's revised position, strongest counter-point and what changed → candidate interpretations (reasoned revision, adopted group line, unchanged and reasoned, unchanged and unexamined) and their basis → individual follow-up at the stated horizon. 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 perspective-taking or reasoning skill without a calibrated model.

Elements and limits

Structured academic controversy, assigned positions, role reversal, perspective-taking, structured discussion, rebuttals, argumentation, sentence starters, consensus building, knowledge synthesis, group roles, peer discussion, rubric, feedback, reflection and assessment. Principles the pattern draws on: Perspective-Taking, Social Learning, Debate, Perspective-Seeking/Multiple Perspectives, Peer Discussion and Purposeful Reflection.

This pattern is scoped to cooperative argument on a question that admits more than one defensible, evidence-bearing answer, with an integrated position as the target. Questions with one correct answer, topics where a two-sided frame would create a false equivalence, and goals of winning or defending a case under challenge (see Debate) need other configurations. The policy supports observation and design reasoning. It remains to be tested whether SAC improves individual argument, perspective-taking, topic understanding or attitudes more than debate or discussion, for which learners and at what horizon. The same is true of whether the reversal or the synthesis carries any effect.

Claims

Examples

  • Students debate a policy question, switch sides, and then produce a joint synthesis.
  • Civics learners examining competing policy responses and then writing a synthesis recommendation.
  • Ethics classes switching sides on a controversial dilemma before seeking common ground.
  • History learners arguing multiple interpretations of a contested event.

Key Sources

  • Johnson, D. W., & Johnson, R. T. (2009). Energizing learning: The instructional power of conflict. Educational Researcher, 38(1), 37-51. https://doi.org/10.3102/0013189X08330540
  • Hess, D. (2009). Controversy in the classroom: The democratic power of discussion. Routledge.
Page metadata
Field Value
type pattern
id structured-academic-controversy
aliases structured-academic-controversy-sac
title Structured Academic Controversy
description A reusable policy for cooperative argument on a contested question, in which pairs argue assigned positions, restate and then reverse them, and write a joint conclusion, expected to support each learner's reasoned, individually checked position, conditional on an arguable question, evidence for both sides, an individual check and a stated horizon; untested as a whole.
status review
grain_size lesson
author Johnson and Johnson
canonical True
generated by claude/unspecified on 2026-10-02
trust unverified — no one has confirmed this page's content
sources 1 — johnson-2009

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