Scaling School Assessment Against External Performance
A school's cohort of 18 students has completed its school assessment. The moderated school marks cluster tightly, with a mean around 72 per cent and a narrow spread, because teachers applied criteria generously. In the external examination, the same cohort's results are lower and much more spread out, with a mean near 60 per cent and a wide range. The school's assessment panel must decide whether to adjust its internal marks before submitting them for moderation, and one teacher argues that because the criteria were applied 'consistently', no change is warranted.
1. Evaluate the teacher's argument that consistency alone justifies leaving the marks unchanged.
Read the worked answer
Consistency is necessary but not sufficient. If the criteria were applied uniformly but generously, the marks are consistently inflated, which is still a validity problem. The evidence here is a clear pattern: the internal mean sits about 12 percentage points above the external mean, and the internal spread is much narrower than the external spread. That pattern suggests the internal tasks were easier, or the standards applied more leniently, or both. Consistency across teachers does not correct a shared upward bias. The panel should therefore review a sample of internal responses against the criteria and external standards before deciding, rather than treating consistency as a defence.
2. What is the risk of simply subtracting a fixed number of marks from every student, and what would a better approach be?
Read the worked answer
A blanket deduction assumes the inflation is uniform, but it is unlikely to be. Students near the top and bottom of the range may be affected differently; a flat adjustment can compress or distort the ranking and misrepresent relative achievement. A better approach is standards-based: re-mark a representative sample of internal assessments against the criteria, identify where judgements were too generous (for example, over-crediting range while under-penalising accuracy), and apply a consistent recalibration informed by that evidence. The aim is to align judgements with the standard, not to mechanically shift numbers. The external results serve as a signal to investigate, not as a formula to copy.