Moderating a Folio of Oral Tasks Before Submission
A Year 12 Vietnamese (Continuers) teacher has nine students. Their school-assessed folio requires four oral interactions recorded across Terms 1–3. Student A's recordings all come from the same 12-minute conversation in Week 2, split into four clips. Student B's clips span three topics but all were recorded in the final week. Student C has one clip where the teacher speaks for 70% of the time. The teacher believes all nine students are 'proficient' and has given A the highest score. The school's moderation panel must decide whether the evidence is comparable and defensible before scores are confirmed. The teacher's marking key lists three criteria: interaction, range, and accuracy, each out of 10, with no descriptors for intermediate bands.
1. Evaluate whether Student A's and Student B's evidence meets the folio requirement. What should the panel recommend for each, and why?
Read the worked answer
Student A - not acceptable. Four clips from one 12-minute conversation are not four interactions. The folio exists to sample performance across time and topics, so one episode gives a single snapshot. Splitting it into clips does not create new evidence: the same task conditions, same interlocutor, same topic, and same preparation apply throughout. The panel should ask for at least two additional recordings on different topics and ideally at different points in the year.
Student B - acceptable in substance, weak in design. Four recordings across three topics satisfy the sampling intent, but clustering them in the final week means the evidence reflects a single developmental point. This is not automatically invalid, but the teacher must justify it (for example illness or timetable disruption) and should note the limitation.
Principle to state: comparability requires that all students be judged on evidence of similar scope, spread, and conditions. A's score, especially the highest in the cohort, is not defensible because the evidence base is thinner than B's despite a higher mark.
Recommendation: withhold confirmation of A's score pending further recordings; accept B's with a documented note; review whether any other student has similarly clustered or single-episode evidence.
2. The teacher's marking key has three criteria out of 10 with no intermediate descriptors. Explain how this undermines the defensibility of the scores and what the panel should require.
Read the worked answer
Problem 1 - no shared meaning between bands. Without descriptors, a mark of 7 can mean 'good range' to one marker and 'adequate range' to another. The panel cannot test whether A's 9 on interaction is justified because there is no statement of what a 9 looks like versus a 7.
Problem 2 - criteria are not independent. 'Interaction', 'range', and 'accuracy' overlap heavily. A student who sustains a conversation will usually show range; a hesitant student may still be accurate. Without descriptors, markers tend to apply a halo effect: one impression drives all three scores. This inflates strong students and depresses anxious ones.
Problem 3 - no anchor evidence. The panel needs annotated samples showing why a performance sits at a given band, not just a number.
What the panel should require: a descriptor set with at least four bands per criterion, written as observable behaviours (for example, 'initiates and extends turns, repairs breakdowns without switching to English'); at least two annotated benchmark recordings per band; and a second marker or moderation sample for the top and bottom scores. The teacher should then re-mark A and C against the descriptors before confirmation.
3. Student C's clip shows the teacher speaking about 70% of the time. What does this indicate about the task design, and how should the panel treat that clip?
Read the worked answer
Diagnosis. A 70/30 teacher-to-student split usually means the task was a question-and-answer check rather than an interaction. If the teacher asks closed questions and the student answers in short turns, the clip measures recognition and short response, not the ability to sustain and manage discourse. It may also reflect a teacher who fills silences, which removes the student's chance to self-repair.
Treatment. The clip should not be counted as strong evidence of interaction. It can support a judgement about accuracy or pronunciation in short turns, but it cannot support a high score on interaction because the student was not given the space to interact. The panel should ask the teacher to re-record with a prompt that requires the student to initiate, extend, and negotiate meaning - for example, a decision-making task with an information gap.
Wider check. The panel should ask whether other students' clips show the same pattern. If several do, the issue is task design across the class, not one student's performance, and all affected scores need review. The teacher should also be advised to keep silent pauses rather than filling them, and to record a sample where the student leads.