Speaking Task: Fluency vs. Accuracy Under a Timer
A candidate records a spoken response to a prompt: 'Describe a time you solved a difficult problem.' She has 45 seconds to prepare and 90 seconds to speak. On her first attempt she spends the full 45 seconds writing a script, then reads it aloud. She finishes in 60 seconds, but her delivery is flat, she stumbles twice reading her own handwriting, and she never states the outcome of the problem. Her pronunciation is generally clear; her grammar is mostly accurate; her vocabulary is adequate. She asks whether she should keep scripting, since 'it makes me more accurate.'
1. Should she keep writing a full script during prep? Evaluate the trade-off using what the task actually rewards.
Read the worked answer
No - a full script is the wrong prep artifact for a 45s/90s timed speaking task.
What the task rewards: fluency, coherence, and communicative effectiveness under time pressure - not written accuracy. A full script optimizes for accuracy (her strongest area already) and costs her in the areas that are actually differentiating:
- Reading aloud ≠ speaking. Flat delivery and stumbles came from transferring written text to voice. Scripts suppress natural intonation and pausing.
- She ran out of content, not time. She spoke 60s of 90s and never stated the outcome - a structure failure, not a language failure. A script didn't prevent it because she scripted sentences, not an arc.
- Prep time is for structure, not prose. 45 seconds is enough to note 3-4 keywords, not to draft and then decode handwriting.
Better prep: write a skeleton - Situation → Problem → What I did → Result - as 4 short keyword lines. Speak from the skeleton. The 'Result' line is mandatory; it's the coherence signal the task is checking.
2. She has 90 seconds and tends to finish early without a result. Give her a concrete structure with rough timings, and explain how it fixes her specific failure.
Read the worked answer
Structure with timings (90s total):
- 0-10s - Situation: one sentence of context. ('In my last job, our weekly report was always late.')
- 10-25s - Problem: what made it hard. ('The data came from three teams and nobody owned the merge.')
- 25-65s - Action: what she did, 2-3 concrete steps. ('I proposed a shared template, set a deadline, and ran a 10-minute check-in.')
- 65-85s - Result: the outcome, ideally quantified. ('Reports went out on time for the next quarter.')
- 85-90s - Close (optional): one line on what she learned.
Why this fixes her failure: her problem was structural, not linguistic - she narrated without landing the outcome. Reserving a fixed time slot for the result makes it impossible to omit; even if the Action section runs long, she can compress it to protect the Result. The keyword skeleton (not a script) keeps delivery natural, addressing the flatness and the reading stumbles. She should rehearse the timing, not the wording.