The Dominant Partner
A paired oral test lasts 12 minutes with two candidates, one examiner and one assessor. On the recording, Nicola produces roughly 70% of the words, opens every new topic, and asks Sofia two questions. Sofia answers four times in exchanges averaging eight words, and takes one 25-second turn when the examiner directs a question at her. The assessor must give each candidate an individual score on five criteria, one of which is interactive communication, and there is no opportunity to re-test. The assessor asks you how to score Sofia fairly.
1. What can you legitimately conclude about Sofia's interactive communication from this sample?
Read the worked answer
Very little about her ability, but something about her performance - and the two must be kept apart. A dominant partner is an artefact of the interaction, not a property of the candidate; the examiner's decision to leave the imbalance unmanaged is a further artefact. Sofia's low output is therefore not evidence of low interactive ability.
But you must still score on the evidence you have, not on a hypothesis about what she might have done. That evidence is thin but non-zero: a 25-second extended turn, four responses to Nicola's questions, and whatever backchannelling and repair behaviour is audible. Score conservatively against the descriptors, cite the specific moments you relied on, and record in the report that the sample was constrained. What you must not do is let the thin sample silently become an ability claim with no note attached.
2. What should the examiner have done during the test, and what options remain now?
Read the worked answer
In a paired format the examiner's job is to protect equality of opportunity. The moment Nicola took a third consecutive topic, the examiner should have nominated Sofia, asked her an open question, or explicitly invited her to respond to Nicola's last point. Twelve minutes is long enough to engineer at least two substantive turns for each candidate.
What remains now:
- Re-listen to the full recording for interactional evidence that a transcript would miss - overlap, backchannels, requests for clarification, responses to Nicola's arguments rather than to the examiner's questions.
- Document the imbalance in the score report, with a timestamp reference, so the result can be reviewed.
- If the stakes are high, add a short examiner-led interaction of two to three minutes, because the original samples are not comparable between the two candidates. If that is impossible, state the limitation and let the awarding body decide - do not quietly inflate or deflate the score to compensate.
3. Nicola initiated, maintained and controlled the interaction. Should her interactive score be in the top band?
Read the worked answer
Not automatically. Initiating and maintaining are only part of the construct; interactive scales also cover inviting others, responding to their contributions, and negotiating meaning jointly. Nicola dominated, made no space for Sofia, and her two questions appear procedural rather than responsive to what Sofia said. On the evidence that she did not build on her partner's contributions, the interactive score should sit below the top band.
Two guardrails against over-correction. First, do not downgrade her on volume alone - volume is not a criterion. Second, some of the failure is the examiner's, and a candidate should not pay for an examiner's failure to manage the room. Base the reduction strictly on what she did not do when the opportunity was there: the two questions that got short answers, the openings where she could have handed over, and any point at which Sofia tried to contribute and was not taken up.