Two Sessions Before the Test
Dana's recorded mock runs 12 minutes. Part 1 answers are 8–12 seconds, accurate but flat. In Part 2 she speaks for 1 minute 40 seconds of the 2 minutes, then stops with 'that's it'. In one two-minute sample she says 'like' 9 times, 'you know' 5 times, 'very good' 4 times and 'interesting' 3 times. In Part 3 she answers in two sentences, one of which restates the question; on one question she says 'I don't know much about that' and stops. Her statements end with rising intonation. Grammar is accurate but limited to simple clauses plus one type of subordinate clause. She has two 45-minute sessions before the test.
1. Which of her features is costing the most, and which is a red herring? Rank and justify.
Read the worked answer
Costing the most: the short, unextended answers in Parts 2 and 3. These are the only features that affect two criteria at once. A 1:40 turn that ends voluntarily and a two-sentence Part 3 answer give the examiner very little language to assess for fluency, coherence or range. The 'I don't know' response is the sharpest case: it removes the assessment opportunity entirely, which is worse than a hesitant answer.
Second: repetitive vague vocabulary ('very good' ×4, 'interesting' ×3). It caps lexical range, but only within one criterion, and the fix is mechanical.
Red herrings:
- The fillers (9 'like' in two minutes) are audible but the message stays coherent; they are a polish item, not the main loss.
- Rising intonation slightly reduces confidence but does not make her unintelligible at this level.
- Grammar is accurate. Drilling complex clauses would risk her fluency for a criterion that is not currently failing. Leave it.
Therefore: both sessions go to extending discourse, with vocabulary substitution as a two-minute add-on.
2. Design the two 45-minute sessions. State what each session trains and what Dana should be able to do at the end that she could not do before.
Read the worked answer
Session 1 - Part 3 extension (45 min)
- 10 min: teach a three-step stall to replace 'I don't know': reframe the question → give a general claim → give one reason. Practise until it is automatic.
- 20 min: ten Part 3 questions, each answered to a fixed shape - claim, reason, example. Hard stop at 45 seconds so she learns to fill the space rather than stop early.
- 15 min: playback. Mark only two things - did she reach the example, and did she avoid restating the question. Exit criterion: five consecutive answers that each contain a reason and an example.
Session 2 - Part 2 length and lexical precision (45 min)
- 15 min: three long turns with a visible timer, forced to keep talking to 1:55 using a prepared structure (what / when / why it mattered / how it changed later).
- 15 min: vocabulary swap drill - replace 'very good', 'interesting', 'nice' with precise alternatives in her own sentences, not a list.
- 15 min: full Part 2 and Part 3 run, recorded, judged against the same two criteria only. Exit criterion: a 1:55 turn with no more than three vague adjectives.
Both sessions end with the same instruction: an answer is not finished when the fact is stated, it is finished when the reason is given.