Why "the student scored 42%" isn't enough data

A percentage is the last thing that happens in an assessment, not the most important one. By the time it lands in a gradebook, almost everything worth knowing about how a student got there has already been discarded.

Which questions were missed, and what they had in common. How long a student spent on each one before answering. Whether the wrong answers cluster around a specific misconception, or scatter randomly. Whether this is the third time this term the student has stumbled on the same underlying skill, dressed up in a different question. All of that is calculable in the moment a test is graded — and almost none of it is kept.

The fix isn't a better-looking report generated after the fact. It's not throwing the inputs away in the first place — mapping every question against the topic, prerequisite, and skill it actually tests, so the signal survives past the moment it's produced. A score answers how much a student got right. Keeping the inputs is what lets you answer why, and what to do about it.