The Silent Partner in a Paired Speaking Task
In a paired speaking task (B1, 4 minutes), Candidate A is dominant and fluent; Candidate B produces short answers, long pauses, and rarely initiates. The examiner must score each candidate separately on interaction, fluency, and accuracy. Two minutes in, A has spoken roughly 80% of the time; B has answered three direct questions with one or two words. B's accuracy on those few utterances is good. The task prompt asks the pair to reach a joint decision about a weekend itinerary. The examiner is tempted to intervene to prompt B, and also tempted to score B low on interaction because he 'did not interact.'
1. Should the examiner intervene to prompt Candidate B? Discuss the trade-off.
Read the worked answer
Intervene only minimally, and only to protect the task - not to coach.
The examiner's job is to elicit a ratable sample, not to run the conversation. If B says nothing for the remaining two minutes, there is no evidence to score, which harms B more than a light prompt would.
A defensible intervention: a neutral, non-leading nudge directed at the pair - 'Remember to reach a decision together' - which shifts responsibility to A to include B, without the examiner feeding B content.
What to avoid: asking B direct questions yourself. That converts a peer-interaction task into an examiner-led interview and changes the construct being measured. It also unfairly advantages B relative to candidates who had to manage a dominant partner unaided.
Record it: note the dominance pattern on the mark sheet so the board understands why B's sample is thin.
2. How should B be scored on interaction and fluency given that A dominated?
Read the worked answer
Score the evidence, but attribute correctly.
Interaction: B did respond appropriately to three direct questions, so he is not at zero. But he did not initiate, negotiate, or contribute to the joint decision - the core of the task. Score at the lower end of the band, while noting that the opportunity to interact was constrained by A.
Fluency: the long pauses and minimal turns are genuine evidence of limited fluency, independent of A's dominance. This band can be scored on what B produced.
Accuracy: good on the few utterances, but the sample is too small to support a high band. Score conservatively and flag 'insufficient sample' on the record.
Fairness principle: penalize B for what he did not do when he had the chance; do not penalize him for turns A never yielded. The mark sheet comment should make this distinction explicit so a reviewer can see the reasoning.