PTE Summarize Group Discussion: Track 3 Speakers and Build a Clear Summary

Track four things while you listen: the shared topic, each speaker’s main position, one useful support point, and the relationship between speakers—then speak from that map instead of replaying the discussion.

Three speakers. The audio starts automatically. Ten seconds to prepare. This is where perfectly decent listening notes can suddenly look like someone dropped a bag of keywords down the stairs. Your goal is not to rescue every word. Build a small map: topic, speaker positions, support, and relationships. That map is what you speak from.

Current task format

Pearson’s current PTE Academic & UKVI Speaking & Writing format says Summarize Group Discussion gives you a discussion between three people and asks you to summarize it in your own words. The prompt can last up to three minutes. After the audio ends, you get 10 seconds to prepare and up to two minutes to speak. You can take notes while the discussion plays.

The task assesses Listening and Speaking. Pearson says the response is scored for Content, Pronunciation, and Oral Fluency. Content is not just “did you mention some correct nouns?” Pearson specifically points to paraphrasing, each speaker’s contribution, relationships between speakers, connected ideas, and beginning with the main topic. Pearson also says Content, Pronunciation, and Oral Fluency are AI-scored traits, with a human expert reviewing Content before the task score is confirmed.

That matters because it tells you what your notes should look like. If your page contains twelve details but no clear speaker ownership or relationships, you have collected information without building a summary.

So the listening job is not “remember everything”; it is “preserve the structure Pearson asks you to report.”

Listen for topic and each speaker’s position

Your first job is separation. Keep the three tabs open without letting them merge.

Suppose a discussion is about whether universities should make attendance at large lectures optional. You hear ideas about recorded lectures, classroom interaction, commuting, and student discipline. Those are details. The topic is the policy question. Each speaker’s position is what they think should happen and why.

A fast note might look like this:

• Topic: optional attendance for large lectures
• S1: supports flexibility — recordings help commuters
• S2: opposes — attendance improves participation/routine
• S3: mixed — optional for some lectures, required for discussion sessions

Notice what is missing: full sentences. You are not a tiny courtroom stenographer. Writing beautifully while the audio keeps moving is a great way to produce elegant notes about Speaker 1 and absolutely nothing about Speaker 3.

When a speaker gives several details, ask one question: What position do these details support? That question prevents vivid examples from stealing more note space than the actual claim.

Note agreement, disagreement, and examples

Once each speaker has a lane, add the lines between the lanes. This is the part many summaries lose.

Use whatever shorthand you can read instantly. For example:

• = agreement
• ≠ disagreement or contrast
• + builds on another point
• ± partial agreement or qualification
• ex example

If Speaker 2 says, “I agree that recordings help students who travel a long way, but I think regular attendance still matters,” do not reduce that to S2 disagrees. The useful relationship is partial agreement + contrast.

Examples should earn their place. Keep an example when it explains the claim, makes a contrast clearer, or gives you one concrete detail to support the summary. Drop it when it is merely the third illustration of a point you already understand.

Lead with the shared topic

Pearson’s Content guidance says stronger responses begin with a summary of the main topic before moving into subtopics and supporting details. That is also the easiest way to stop your response sounding like three unrelated reports.

Compare these openings:

Weak structure: “Speaker one said recorded lectures are useful. Speaker two said attendance is important.”

Stronger structure: “The discussion focused on whether attendance at large university lectures should remain compulsory, with the speakers balancing flexibility against the benefits of regular participation.”

The second opening gives the listener a container. Now each speaker’s contribution has somewhere to go.

You do not need a grand template. Useful openings are short:

• “The discussion focused on…”
• “The speakers discussed whether…”
• “The group considered the advantages and drawbacks of…”

Choose the one that fits the actual discussion. A memorized twenty-second introduction is expensive decoration when your answer time is limited.

Attribute without a transcript dump

Attribution matters because the task is about a discussion, not a bag of anonymous ideas. But repeating “Speaker 1 said… Speaker 2 said… Speaker 3 said…” for every sentence can sound mechanical.

Vary the reporting structure while keeping ownership clear:

• “The first speaker argued that…”
• “The second speaker agreed with the concern about cost but added that…”
• “In contrast, the third speaker preferred…”
• “Both the first and third speakers emphasized…”

The goal is not fancy vocabulary. It is efficient attribution.

Synthesize relationships

Attribution tells us who owns an idea. Synthesis tells us how the ideas fit together.

Useful relationship language includes:

• Agreement: “Both speakers agreed that…”
• Partial agreement: “The second speaker accepted X but argued that…”
• Contrast: “In contrast, the third speaker…”
• Build-on: “Building on that point, the second speaker added…”
• Shared concern: “Although their solutions differed, all three were concerned about…”

This is where the three open tabs finally become one map.

Do not invent relationships that were not present. If Speaker 3 introduces a new issue, do not force “however” simply because contrast words sound academic. Connect the ideas according to meaning.

Production practice: Take these notes and say two connected sentences aloud:

• S1: online office hours increase access
• S2: agrees access better; worries questions become less spontaneous
• S3: suggests hybrid office hours

Use original wording

“Use your own words” does not mean “replace every normal verb with something that sounds like it escaped from a nineteenth-century committee report.” It means preserve the meaning while changing the wording and structure naturally.

Focus on safe paraphrase moves:

• change the sentence structure;
• replace a phrase with a genuine synonym;
• compress examples into their function;
• combine related points;
• report stance with verbs such as argued, suggested, preferred, questioned, agreed, or emphasized when they match the speaker’s meaning.

A useful collocation check: speakers raise a concern, give an example, make a point, support an argument, reach agreement, or take a different view. Small collocation choices help a summary sound natural without turning it into a vocabulary exhibition.

A note-taking grid

Here is the reusable control panel. Keep it compact enough that you can actually use it while listening.

Topic: ______________________________

Do not turn each cell into a paragraph. If the grid becomes a browser with 37 tabs and music playing somewhere, it has stopped doing its job.

A practical order while listening is:

1. Write the topic as soon as it becomes clear.
2. Give each speaker one main-claim phrase.
3. Add only one support detail when it helps explain the claim.
4. Mark relationships as soon as you hear them.
5. Use the 10-second preparation window to choose your opening and speaking order, not to rewrite everything.

A full practice walkthrough

The discussion below is original practice material written for this article. Read it once at a natural pace, or ask a study partner to read the three roles aloud. On your first pass, do not look at the model grid.

Practice discussion: Should universities record every lecture?

Speaker 1: I think universities should record all large lectures and make them available to students. People miss classes for legitimate reasons, such as illness or transport problems, and recordings also let students review difficult explanations before exams. I’m not saying students should stop attending, but having the recording gives them a useful backup.

Speaker 2: I agree that recordings can help when someone misses a class, but I’m not convinced every lecture should be recorded automatically. If students know the video will always be there, some may postpone the lecture and fall behind. There is also a difference between watching later and being present when the lecturer asks questions or checks whether the class understands.

Speaker 3: I can see both sides. For very large lectures, recordings seem practical because interaction is already limited and students may need to revisit technical material. But in smaller classes, especially seminars, recording everything could reduce participation because students may feel less comfortable asking questions. I’d use recordings selectively rather than applying one rule to every class.

Your turn: build the grid first

Write the shared topic, then fill S1, S2, and S3 with one claim, one support point, and a relationship. After that, give your own spoken summary before revealing the model.

The model is not a script to memorize. Its job is to show the transformation: three separate tabs while listening, then one connected map while speaking.

Now do a second attempt from the same notes without reading the model. Change the wording, keep the meaning, and aim for clean phrasing rather than maximum speed.

Self-review

After your response, review the same things the task actually asks you to do. Do not turn self-review into a mysterious score prediction.

If Content is weak, return to the grid. If your notes are good but the delivery is hesitant, practise the same summary again instead of adding more facts. The problem may be output, not comprehension.

Reuse the original discussion above for one more spoken summary, then practise another spoken response with FunFluen by choosing a general English speaking-practice path. This is general speaking practice; the exact PTE multi-speaker task is not claimed to be preloaded or scored there.

For the broader exam family, return to PTE Speaking.

You do not need a transcript in your head. You need three clean speaker tabs and the lines between them. Once those survive the listening, the microphone is no longer asking for a memory miracle. It is asking you to speak from a map you already built.

Sources

• Pearson PTE — PTE Academic & UKVI test format: Speaking & Writing: current task format, timing, note-taking, Content guidance, Pronunciation, Oral Fluency, and human Content review.
• Pearson PTE — Mastering the ‘summarize group discussion’ task for PTE Academic success: task-specific preparation context on active listening, note-taking, and coherent multi-perspective summaries.