The listening section runs around 47 items in about 29 minutes across four task types, and they split into two kinds that need opposite things from you.
The fast ones: Listen and Choose a Response. Short conversational exchanges where you pick the natural reply. Note-taking is useless here — by the time you have written anything the item has moved. What works is staying alert between items and answering on instinct, because the response that sounds right usually is.
The functional ones: Listen to an Announcement, Listen to a Conversation. Campus and everyday material — a schedule change, an enquiry, an exchange between two people. Hold three things: who is speaking, what they want, and what is decided. Those three answer most of what is asked.
The long one: Listen to an Academic Talk. Here notes earn their keep, in a fixed shape rather than as transcription:
- the topic, in three words
- three or four points, one line each
- the conclusion or the speaker's position
Trying to write everything means understanding nothing, and on a lecture the questions are about the argument rather than the details anyway.
Accents are part of the test. ETS uses a range of native-speaker accents. This is not something to study; it is something to be exposed to. Ten minutes of English podcast or news audio a day for three weeks does more for this section than any strategy.
Recover fast. With items this dense, missing one is normal. Candidates lose far more to dwelling on a lost item — and missing the next two while they do it — than to the item itself.