Scaling and Comparing School-assessed Scores Across Classes

Why not comparable. A mark is only meaningful relative to the task and the marking standards applied. X's task was harder and X penalised accuracy heavily, so a given mark represents a higher level of performance than the same mark under Y. Y's familiar topic and effort credit inflate marks. The narrower spread in Y (68–88) suggests either a more homogeneous class or a compressed marking scale; the wider spread in X (55–92) suggests a discriminating task.

Consequence. Comparing 89 from X with 85 from Y as if they were the same unit is invalid. Rank within class is not comparable across classes either, because the class composition differs.

Evidence needed before adjustment:

1. The task descriptions and marking keys from both classes.
2. Annotated samples from each band in each class.
3. A common moderation task - ideally the same stimulus and prompt given to both classes, marked blind by both teachers.
4. Class composition data (prior achievement, language background) to judge whether the mean difference reflects teaching or intake.

Without this, any adjustment is guesswork dressed as fairness.

Step 1 - Shared descriptors. Both teachers agree band descriptors before marking. This is the foundation; without it, later steps are cosmetic.

Step 2 - Common task. Both classes complete one identical assessment under identical conditions. This creates a genuine anchor point.

Step 3 - Blind cross-marking. Each teacher marks the other's sample of six scripts (top two, middle two, bottom two) without seeing names or class. Discrepancies of more than one band trigger discussion.

Step 4 - Statistical check, used cautiously. Compare means and spreads on the common task only. If X's class mean is substantially lower on the same task, that is evidence of marking severity, not of weaker students. Do not mechanically curve to equalise means; equal means can hide real differences in achievement.

Step 5 - Documented adjustment. If severity is established, adjust X's marks upward by a stated amount applied consistently, and record the reason. Avoid adjusting individual students selectively.

Communication to parents. Explain that marks were checked against a common task and shared standards, that adjustments were applied to a whole class rather than to individuals, and that the aim is that a given mark means the same thing in both classes. Provide the descriptors so the claim is checkable.

Acknowledge the concern. The student's performance has not changed. Any adjustment is about the meaning of the mark, not about the student's effort or ability.

Explain the logic. If Y's task was easier and effort was credited, then an 85 in Y may represent a lower standard of performance than an 85 in X. Leaving both unchanged means two students with different achievement receive the same number, which disadvantages the stronger performer in X and misleads universities and future teachers about both students.

Show the evidence. Present the common task results and the blind cross-marking. If the Y student performs at the same level as the X student on the common task, no downward adjustment is warranted - the concern is resolved by evidence. If the Y student performs lower, the adjustment is justified.

Limit the harm. Apply adjustments to the class cohort consistently rather than singling out individuals, and avoid adjustments that push students across a grade boundary without clear evidence.

Be transparent about uncertainty. If the evidence is thin, prefer no adjustment and a documented note over a confident re-ranking that cannot be defended. Fairness here means consistent treatment, not identical numbers.