School-assessed oral: balancing fluency against accuracy
You are moderating a SACE Stage 2 Korean (background speakers) school assessment. Two teachers have rated the same student's 6-minute individual oral presentation (topic: 세대 갈등, generation conflict). Teacher A gives 17/20, praising fluent delivery, natural intonation, and confident use of connective forms such as -기는 하지만 and -더라도. Teacher B gives 13/20, noting that the student used honorific -시- inconsistently when referring to grandparents, said 할아버지께 물어봤어요 once but 할아버지한테 물어봤어요 twice, and made two tense slips in past narration. The student is a heritage speaker who speaks Korean at home but has had little formal instruction. The school's task sheet weights 'interaction and fluency' and 'grammatical accuracy and range' as separate criteria, each out of 10. The panel must reconcile the two scores before submitting to the SACE Board.
1. Identify the main assessment error suggested by the 4-point gap, and explain how the separate criteria should be applied.
Read the worked answer
The gap points to criterion conflation: Teacher A is letting fluency halo the accuracy judgement, while Teacher B is letting accuracy errors drag down the fluency judgement. The fix is to score each criterion independently.
- Fluency/interaction (out of 10): the evidence described (natural intonation, sustained 6-minute delivery, confident connectives) supports a high score, plausibly 8–9. Heritage fluency is legitimate evidence; it is not discounted because the student lacks formal instruction.
- Accuracy/range (out of 10): honorific inconsistency with grandparents and past-tense slips are real, describable errors. But the descriptor asks about range as well as accuracy, and the student shows range (-기는 하지만, -더라도). A mid-to-high score such as 6–7 is defensible.
A reconciled total near 14–16/20 is more consistent than either 17 or 13, because it reflects two distinct judgements rather than one impression.
2. Teacher B argues the honorific errors are 'serious because Korean honorifics are culturally fundamental.' Evaluate this reasoning against the task's criteria.
Read the worked answer
The claim is partly right but misapplied. Honorific misuse is a genuine accuracy error and should be named in feedback, since -시- and particle choice (께/한테) are core to formal register. However, three cautions apply:
- Weighting: the task sheet does not assign honorifics special double weight. Treating one error type as decisive distorts the criterion.
- Context: the oral is a presentation, not a formal letter; inconsistent honorifics in spontaneous speech are common even among educated speakers and should be judged against the descriptor's band, not against an idealized native norm.
- Consistency: if honorifics are penalized heavily, the same standard must apply to all candidates, including non-heritage students. Otherwise moderation becomes inequitable.
The defensible position: note the errors clearly, deduct within the accuracy criterion proportionally, and do not let them override fluency evidence.
3. Recommend how the panel should document the reconciled score so it survives external moderation.
Read the worked answer
Documentation should make the reasoning auditable without re-scoring by impression:
- Criterion-by-criterion annotations quoting specific evidence (e.g. 'fluency: sustained 6 min, natural intonation, effective -더라도'; 'accuracy: honorific particle inconsistency ×2, past-tense slips ×2').
- A short rationale explaining why the final score sits between the two initial judgements, referencing the descriptor bands.
- Retain the recording (with consent) so a moderator can verify claims.
- Note the heritage context factually, not as a bonus or penalty.
This protects against the common moderation finding that a school's marks are internally inconsistent: the record must show the score followed the evidence, not the teacher's overall impression.