The Ambiguous Listening Item
You are writing a Listening Comprehension item for a B2 test. The audio is a 40-second voicemail: a manager says, 'I'd love to have the report by Friday, but honestly, whenever you can manage is fine - just don't leave it till Monday.' The intended answer is 'Friday.' A pilot with 30 learners shows a 50/50 split: half chose 'Friday', half chose 'whenever.' The item stem reads: 'When does the manager want the report?' A reviewer says the item is fine because 'Friday' is stated first.
1. Is the reviewer correct that the item is valid because 'Friday' is stated first? Explain the listening construct at stake.
Read the worked answer
No. The reviewer confuses first mention with intended meaning.
What the audio actually encodes:
- 'I'd love ... by Friday' = preference/ideal.
- 'but honestly, whenever you can manage is fine' = concession that overrides the preference.
- 'just don't leave it till Monday' = the only hard constraint (a deadline, not Friday).
A 50/50 split is a red flag, not noise: it means the item is testing pragmatic inference of concession, and the stem 'want' is ambiguous between 'prefer' and 'require.' The 'Friday' answer rewards surface listening; the 'whenever' answer rewards deeper inference. Both are defensible, which makes the item unreliable.
Fix: either rewrite the stem to 'What is the manager's actual deadline?' (answer: before Monday), or edit the audio so the concession is unambiguous.
2. Rewrite the item so it has a single defensible answer, and state what skill it now measures.
Read the worked answer
Revised stem: 'What is the one thing the manager insists on?' Options: A) The report arrives by Friday. B) The report arrives before Monday. C) The report can arrive any time. D) The report is not needed. Key: B.
Why this works:
- 'Insists on' targets the hard constraint ('don't leave it till Monday'), which is the only non-negotiable element.
- Distractor A is the surface 'Friday' preference - now clearly a distractor, not a competing truth.
- Distractor C collapses the concession into 'any time', ignoring the Monday limit.
Skill measured: recognising concession and distinguishing preference from obligation in spoken discourse - a genuine B2 listening sub-skill, not recall of a single word.