TOEIC Link Part 3: Keeping Three Speakers Straight Without Losing the Thread

Most Part 3 conversations have two speakers, and your ear settles into the back-and-forth quickly. Then a three-speaker item arrives and the rhythm breaks. A third voice enters, the turns stop alternating predictably, and the questions — which on these items almost always ask who said or wanted what — suddenly require you to have tracked attribution the whole way through. Candidates who cruise through two-speaker items stall here, not because the language is harder but because the bookkeeping is. This guide is about carrying three speakers through a conversation without dropping the thread.

Why the third speaker changes everything

With two speakers, attribution is nearly free: if it was not one, it was the other. Your brain tracks the conversation by content and lets speaker identity ride along. Add a third voice and that shortcut dies. Now "who suggested the change" has three possible owners, and the answer choices are built to punish a vague memory. The test knows that two-speaker tracking is automatic and three-speaker tracking is not, so the three-speaker item is where it places the attribution questions that discriminate.

The load also rises because three speakers usually means a more complex situation — a meeting, a customer with two staff, a handoff between departments. More roles, more competing wants, more chances to mix up who holds which position. The comprehension is not deeper; the tracking is heavier, and tracking is a skill you can drill.

Anchor each speaker by role before the content arrives

The single most useful move is to fix each speaker's role in the first few seconds, before the substance starts. The conversation almost always establishes who everyone is early — a manager and two team members, a guest and two hosts, a buyer and two sellers. Catch those roles immediately and you have a frame to hang every later statement on.

Roles matter more than names, and chasing names is a common mistake. Part 3 questions ask "what does the woman suggest" or "what will the man in sales do," not "what does Mr. Tanaka want." You rarely need the name; you always need the role and the gender, because those are what the questions and answer choices key on. Listen for "as the project lead" or "from the design side," not for proper nouns. This role-first habit is the listening counterpart to reading speaker attitude and tone from limited cues — you are building a model of each person before the details land.

Track positions, not just lines

Once roles are anchored, track what each speaker wants rather than transcribing what each one says. Three-speaker items typically set up a small disagreement or a division of labour: one person proposes, one objects, one resolves. If you hold each speaker's position — proposer, objector, decider — you can answer the attribution questions without having memorised the exact words.

This is where light notes earn their cost. You do not have time to write sentences, but two or three letters per speaker — a tag for the role and a mark for their stance — keeps the model from collapsing under the second and third voice. The goal is a skeleton you can read off when the questions come, the same minimal-note discipline that supports extracting action items and decision points from any meeting-style audio.

Use the questions to pre-aim your attention

You see the three questions before the audio plays — on three-speaker items, use that preview harder than usual. If one question asks "what does the woman recommend" and another asks "what will the men do next," you know in advance that attribution is the test and that you must keep the woman's recommendation and the men's plan separate. Pre-aiming tells you which threads to guard, so the third speaker does not pull your attention off the line that the questions actually reward.

This pre-aiming also rescues you when the audio gets dense. If you can only hold two of three threads cleanly, the question preview tells you which two to prioritise — protect the threads the questions ask about and let the incidental chatter wash past.

Drill attribution under load, not comprehension

Most Part 3 practice tests whether you understood the conversation, which two-speaker items already train. To prepare specifically for three-speaker items, change what you drill: after listening once, before checking answers, write down who held each position — who proposed, who objected, who decided — and only then look at the questions. If you can reconstruct the attribution map from memory, the questions are trivial; if you cannot, you have found exactly the skill the three-speaker item is testing. Practising the map, not the gist, is what turns the dreaded three-speaker conversation into one of the more predictable items on the section, leaving you composed for the graphic-linked questions where attention is split a different way.