Fishbowl Discussion: A Classroom Strategy for Better Reasons

A Fishbowl discussion is a talk routine in which a small inner circle discusses a question while a larger outer circle listens with one clear job. Observers then feed back a reason that moved the talk on, the circles swap, and the second group discusses again to improve the claim.

Fishbowl at a glance

Fishbowl is an Exploratory talk routine from the Group-Based Discussion set. It takes 8 to 20 minutes and uses an inner and an outer circle. Its purpose is to help learners observe talk and revise ideas. Choose it when your class needs to notice which reasons actually move a discussion forward.

• Family: Exploratory talk, with Group-Based Discussion as the secondary set.
• Time: 8 to 20 minutes, including a debrief and a second round.
• Layout: a small inner circle that talks and a larger outer circle that listens.
• Listener job: Note one reason that moved the talk on.
• Sentence stems: “Our reason is…” and “I’d build on…”
• Check: after the swap, every learner says or writes how the claim changed.

New to the wider field? Start with what we mean by oracy, then browse the Oracy A to Z for neighbouring routines.

When to choose Fishbowl, Socratic questioning or Harkness

Choose Fishbowl when you want part of the class to watch reasoning closely and then use what they noticed. Choose Socratic questioning when you want to probe one line of thinking through teacher questions. Choose a Harkness discussion when the whole group should share responsibility for a long, evidence-heavy conversation around one table.

The difference is who does the noticing. In Socratic questioning, a questioner asks probing questions such as “what do you mean?” and “how do you know?”. The person asking therefore does most of the listening. Fishbowl hands the noticing to the outer circle, and their report shapes the next round.

Harkness discussion usually seats the whole group around one table with a shared text while the teacher steps back. It suits groups who can already sustain talk without a split structure. Fishbowl is more scaffolded. It has fewer voices at once, a defined listening job and a planned second attempt.

Fishbowl is the wrong choice when every learner needs to talk at once. Use Think-Pair-Share or Turn and Talk for quick rehearsal. Use P4C for a community enquiry into a philosophical question. Our class discussion techniques cover habits that help any format. Fishbowl stays exploratory talk even when speakers defend a position, because the aim is to test and revise thinking, not to win.

How to run a Fishbowl in your next lesson

Set one question, one shared piece of evidence and one listening job before anyone moves a chair. Run a short first round, take brief feedback on a single useful reason, then swap the circles so the second group improves the claim. Finish with an individual check so every learner shows their own thinking.

Before the first round

Display the question, the evidence and the listener job. Choose four to six speakers, and ask everyone else to observe with a card. Agree two or three discussion guidelines, such as “respond to the last reason before adding a new one”. Then brief observers precisely: “Note one reason that moved the talk on. Not the loudest point or the longest. The reason that made the group think again.” If you use Talk Tactics roles, keep them light.

Run the round and debrief

Give the inner circle four to six minutes. Step in only if the talk loses the evidence: “Which part of the source supports that?” Then ask three or four observers to report with “Our reason is…”. Each report should name a reason and say what it changed. Feedback reports a useful reason. It never scores a speaker’s confidence, fluency or accent. Write the reported reasons on the board as the brief for round two.

Swap and check

Swap circles. The new inner circle starts from the board, not from scratch: “I’d build on…”. Their task is to improve the claim, not repeat it. Afterwards, every learner writes or says one sentence: “The claim changed from… to… because…”. This individual check matters, because a lively second round can hide learners whose thinking has not shifted at all. The written version also gives you a record.

Observation card

Give each observer a card with three boxes. Three boxes stop observers writing transcripts and keep their attention on the reason that mattered. The example below is illustrative and written for the maths Fishbowl in the next section. Adapt the wording for your subject. You can also prepare printable cards with Say It.

Worked example: a maths Fishbowl with a better second round

In this maths example, Year 7 learners test the claim “multiplying always makes a number bigger”. The first round produces examples but no rule. The debrief surfaces one reason that broke the claim. The second round turns that reason into a conditional claim, which every learner then writes down for themselves.

Question on the board: “Does multiplying always make a number bigger? Test it with whole numbers and fractions.”

Evidence: four calculations on the board: 6 × 3, 6 × 1, 6 × ½ and 6 × 0.25.

Round one (inner circle A): Speakers agree quickly that 6 × 3 = 18. One says, “Our reason is that times tables always go up.” Another works out 6 × ½ = 3 and hesitates: “That went down.” The group calls it an exception and moves on. They end with “multiplying usually makes it bigger”.

Debrief: Most observer notes record the 6 × ½ moment. One observer reports, “The reason that moved the talk on was 6 × ½, because it showed ‘always’ was wrong, but nobody explained why.” The teacher adds two prompts to the board: “What is special about ½?” and “What happens with 6 × 1?”

Round two (inner circle B): The new speakers start from the board. “I’d build on the half example. Multiplying by ½ means taking half of 6, so it has to be smaller.” Another adds, “And 6 × 1 stays as 6, so 1 is the boundary.” They test 6 × 0.25 and agree it is smaller again.

Improved claim: “Multiplying a positive number by something greater than 1 makes it bigger. Multiplying by exactly 1 leaves it the same. Multiplying by something between 0 and 1 makes it smaller.”

What improved: the claim gained conditions and named a boundary case. It also now rests on a reason rather than a pattern: multiplying by a fraction takes part of the number. Negative numbers make a good opening question for the next lesson.

Individual check: each learner completes “The claim changed from… to… because…”. Look for a reason about the size of the multiplier, not just “we found an example that broke it”.

Make this for my lesson

Fishbowl prompts for maths, science, English, history and geography

Every subject version follows the same shape. There is shared evidence and a question that needs a reasoned answer. Observers note the reason that moved the talk on, and the second round must improve the claim. The table gives a ready prompt for each subject; swap in your own topic and source.

Give English and history sources before the round, or read them together, so the inner circle is not decoding cold. In science, keep both explanations visible throughout. In geography, ask round two to name who gains and who loses. The oracy and critical thinking hub and our discussion techniques guide offer further structures.

Making Fishbowl work for SEND and EAL learners

Make the structure visible and keep the stakes low. Display the question, the evidence and the listener job. Let learners prepare a written point before speaking, and offer a seated layout if moving chairs is impractical. Nobody should have to perform publicly to take part, and feedback should concern reasons, never delivery.

• Layout: the circles are roles, not furniture. Seat the inner group at the front tables and keep observers at their desks, or use two groups on opposite sides of the room.
• Visible supports: keep the question, source, listener job and both stems on screen throughout.
• Preparation: allow a minute of silent notes before round one, or a quick Think-Pair-Share rehearsal.
• Written routes in: a learner can hand a written point to a peer to read aloud, or contribute an observer note before a supported swap.
• EAL: pre-teach key vocabulary from the source, model each stem in a full sentence and offer optional rehearsal with a partner.
• Agreed plans: follow existing communication plans and adjustments. Check with your SENCo if you are unsure what helps a particular learner.
• Feedback: add “we comment on reasons, not on how someone speaks” to your discussion guidelines.

Treat these as options. Ask learners afterwards what made it easier to take part, and adjust next time.

What the evidence does and does not say

There is no Fishbowl-specific effect size cited in this guide, and none should be attached to it. What can be said is modest: Fishbowl gives learners a structured opportunity to hear, examine and revise reasons. Whether that happens in your room is something you can see in the second round.

The Education Endowment Foundation toolkit entries on oral language interventions and collaborative learning approaches summarise research on broad families of practice. Their average estimates describe those families, which include many different programmes. They do not show that an inner and outer circle improves attainment. For that reason, their figures are not presented here as Fishbowl results.

The routine relies on a simple mechanism. Observers listen during the talk, and the debrief uses what they heard to improve the next round. To judge whether that happened, compare the round one and round two claims. Did the second claim gain a condition, evidence or a clearer reason? Did quieter learners contribute? Do the individual sentences show each learner’s own shift? Those checks, tied to your wider aims for oracy, tell you more than a borrowed figure.

Fishbowl discussion: frequently asked questions

These answers cover practical questions about running a first Fishbowl: group size, timing, quiet learners, how it differs from debate and what to do without space to move. Each answer assumes the three-step card and the listener job described above, so you can adapt them quickly for your class.

References and further reading

The sources below give wider context for the evidence card. They are summaries of broad oral language and collaborative learning research, not studies of Fishbowl itself. The internal guides extend the routines this article compares with Fishbowl and help you place it within a whole-school approach to talk.

• Education Endowment Foundation. Teaching and Learning Toolkit: Oral language interventions.
• Education Endowment Foundation. Teaching and Learning Toolkit: Collaborative learning approaches.

Further reading on Structural Learning:

• Oracy A to Z
• Exploratory talk
• Socratic teaching techniques
• Oracy and critical thinking hub
• Say It oracy toolkit