Listening Comprehension: Handling Repair, Paraphrase and Numbers
What the listening domain tests
Listening comprehension uses short monologues and dialogues - announcements, interviews, phone calls, short talks - at near-natural speed, in a mix of British, American and other accents. Typical tasks are multiple choice, matching speakers to statements, and note or sentence completion under a strict word limit. Each recording is normally heard twice, which makes the first pass a gist pass and the second a confirmation pass, not a fresh start.
Rules that decide marks
- Use the pause before the recording to predict what kind of answer is coming: a number, a time, a place, a reason, a name?
- First listening: structure - who is speaking, what they want, how many points are made. Second listening: detail and confirmation.
- Answers are paraphrased. If an option repeats the exact words you hear, treat it as suspect until the meaning confirms it.
- Speakers repair themselves: originally... but, sorry, I mean, actually. The correction is the answer.
- Write shorthand while the audio runs and transfer answers afterwards; do not compose full sentences while someone is still talking.
- Numbers, dates and spellings are tested deliberately: 13/30, 15/50, the 1980s, double letters in surnames.
- If you lose one item, abandon it and re-locate using the next question. The recording will not wait.
Worked example
Recording: The workshop was going to be on the fifteenth of March, but we've had to move it - sorry, the thirteenth of April, not the thirtieth.
Question: When is the workshop?
Reasoning: was going to be cancels the first date. Sorry marks a self-repair, and not the thirtieth explicitly disposes of the near-sounding alternative. Answer: 13 April. A candidate who writes 30 April heard the number but missed the negation; one who writes 15 March stopped listening after the first clause.
Traps
- Writing down the first number, date or name you hear.
- Missing a negative: not included, no longer available, rather than, except.
- Answering with words from the audio that break the word limit or the grammar of the gap (costs 20 pound).
- Losing concentration after one missed item and therefore missing the next two.
- Confusing similar-sounding pairs under connected speech: can/can't, thirteen/thirty, fifteen/fifty.
Exam audio moves on regardless of you; the discipline of skipping a lost item is worth more than any single answer.