IELTS Listening Part 3: Academic Discussion (2026)

Part 3 is where IELTS Listening stops being about facts and starts being about people. Two or three speakers — usually students, sometimes with a tutor — discuss coursework: planning a presentation, reviewing research, debating design choices. The vocabulary is academic but not technical; the difficulty is social. Everyone sounds reasonable, everyone half-agrees, and the options all get mentioned.

Your job is not to understand every idea. It is to answer one question per item: who thinks what, and do they agree?

For context, see the IELTS Listening guide and the question types breakdown.

What Part 3 Tests

Part 3 is a discussion in an educational or training context: seminar preparation, project feedback, library research planning, field-trip debriefs.

It tests whether you can:

• Distinguish speakers and attribute opinions correctly
• Detect agreement, disagreement, and qualified agreement ("yes, but…")
• Follow an argument across turns — the answer often emerges over 3–4 exchanges, not one line
• Handle academic paraphrase: options restate ideas in different words

Format: 10 questions in two sets, commonly matching (e.g. match each student to an opinion) plus multiple choice (single answer or "choose TWO"). The speakers' turns come in order, but options within a matching set are discussed flexibly — track stems, not letters.

Scores here separate Band 7 from Band 8. Protect easy marks in Parts 1–2 first (see Band 9 Listening), then invest practice time here and in Part 4.

The Part 3 Method: Roles, Stances, Elimination

1. Assign roles in pre-reading time

Before the audio, label the speakers functionally:

• Student A / Student B (or named: Mia / Jonas) — the opinion-holders
• Tutor — usually asks questions and summarises; rarely the answer-holder in matching tasks

Read the stems, not just the options. If the stem says "Whose opinion is this?", your listening target is speaker identity. If it says "What do they decide about X?", your target is the agreed outcome — often the last position stated.

Underline differences between options. In "choose TWO from five" lists, options differ by one concept each — circle those distinguishing words.

2. Track stances, not sentences

From the first exchange, jot a micro-stance per speaker: "M: pro-fieldwork, J: prefers lab data". Update it as they concede or refine. Most Part 3 answers reduce to:

• Who proposed it → who supported/opposed it → what they finally agreed

Listen for agreement language (your real answer key):

• Full agreement: "exactly", "absolutely", "I couldn't agree more", "that's just what I thought"
• Qualified agreement: "up to a point", "I see what you mean, but…", "that's true, although…"
• Disagreement (often soft): "I'm not so sure", "I take your point, but…", "haven't we already…?", "wouldn't it be better to…?"
• Decision markers: "so we're agreed…", "let's go with…", "we'll settle on…"

Illustrative snippet (invented example, not a real test): "I see your point about interviews, but the survey data covers twice as many people — shall we lead with that?" The pattern — concession + counter-reason + proposal — is the classic Part 3 decision shape.

3. Eliminate aggressively

Every option will be voiced. The wrong ones are:

• Mentioned then rejected ("we could do X, but there's no time")
• One speaker's view attributed to another (the stem names Mia; Jonas said it)
• True but not the answer to this stem (an advantage mentioned when the question asks for a drawback)

Cross out dead options live. In "choose TWO" tasks, finding three definite rejects is as good as finding two answers.

Deeper option tactics live in question types; weekly planning in the practice guide.

Traps in Part 3

1. Speaker confusion. Two students of the same gender/pitch agree halfway through and you lose track of who said what. Fix: anchor on names/titles ("Mia", "the tutor") and re-anchor every time a new question stem names someone.

2. First-mention fallacy. The first opinion you hear feels like the answer. In discussions it is usually the starting position, later modified. Wait for the resolution — decisions come at the end of an exchange.

3. Paraphrase gaps. Option: "data collection was too narrow." Audio: "we only spoke to first-years, so it's hardly representative." No shared words, same meaning. Pre-reading for concepts (sample, scope, method) beats listening for keywords.

4. Tutor summaries. Tutors often restate both views neutrally ("so one option is X, the other is Y") before students decide. Do not mistake the summary for the decision — the answer is what they choose, usually signalled right after.

5. Polite disagreement sounding like agreement. "That's an interesting idea…" plus a pause often precedes rejection. Listen to what follows the politeness, not the politeness itself.

Practice Drill: The 30-Minute Discussion Session

1. 12 min — timed Part 3 set. Pre-label speaker roles; jot one-word stances per question. No pausing.
2. 10 min — transcript autopsy. For each miss, classify: speaker mix-up, paraphrase miss, premature pick, agreement misread. Re-listen to those exchanges only.
3. 8 min — agreement-phrase harvest. Copy every agreement/disagreement marker from the transcript into a running list. Aim for 30+ phrases in your notebook within a month.

Rotate focus weekly: matching-heavy sets one session, multiple-choice-heavy the next. Track accuracy separately — most candidates are weaker at one. Structure the week with the practice guide and calibrate targets with Band 9 Listening.

Quick-Reference: Discussion Moves Table

Discussion move	Language clues	What it means for answers
Proposal	"why don't we…", "how about…", "we could…"	Candidate answer — wait for reaction
Support	"exactly", "good point", "that would work"	Likely agreed outcome
Soft objection	"I take your point, but…", "up to a point…"	Answer is the compromise, not the proposal
Rejection	"the problem is…", "there's no time for…", "that's already been ruled out"	Eliminate this option
Decision	"so we'll…", "let's settle on…", "agreed?"	Final answer — mark it and move on

Print this mentally: proposal → reaction → decision. The answer is almost always at the decision stage.

Worked Example: How a Decision Question Unfolds

Take a typical matching item: "Who thinks the survey sample was too small?" The exchange runs in layers. First Jonas presents results confidently. Then Mia raises a doubt — "we only asked first-years, didn't we?" Jonas concedes partially — "true, but the trends are still clear." The tutor asks what they'd do differently. Mia proposes widening the sample; Jonas agrees, "fair enough, let's note that as a limitation." The answer is Mia, but it only becomes certain at that final concession. A candidate who picked Jonas at the first confident sentence fell for the starting-position trap.

The defence is patience plus shorthand. When Jonas speaks confidently, jot "J: confident" — not a final answer, just a stance. When Mia doubts, jot "M: sample narrow". When Jonas concedes, upgrade to "agreed: limitation". Your pencil holds the evolving state so your memory doesn't have to. One extra habit helps: note the tutor's questions too, since they preview the next decision point ("what would you change?" signals an upcoming agreement). In "choose TWO" tasks the same patience pays twice: the first correct option often arrives early as a proposal, the second only after objections are resolved. Wait for both resolutions before committing, and use the elimination column — every rejected option you cross out narrows the final pick and protects you from second-guessing..

The Bottom Line

Part 3 rewards speaker-tracking over vocabulary. Assign roles, listen for agreement, eliminate relentlessly, and let decisions come to you.

Next: sharpen option skills in question types, then build discussion practice into your week with the practice guide.