Comparing a semester of writing at once
One score is a data point. A semester of them from the same student is a baseline you didn’t have to build separately.
Save every report, not just the flagged ones
A folder of a student’s scores across the term, including the unremarkable ones, is the baseline you’ll wish you had the first time one comes back high. Building it after the fact means comparing against memory instead of records.
Watch for drift, not just spikes
A gradual rise across several assignments is a different signal than one anomalous score in an otherwise steady pattern — the first might reflect a real change in how a student is working, worth a conversation on its own terms rather than treated as one-off.
A consistent pattern is itself informative
A student who scores in a similar range across a whole semester, flagged or not, is probably just a writer with that rhythm. That consistency is evidence toward reading a single mid-range score as ordinary for them, not toward assuming the worst.
Set this up at the start of term
Deciding to keep every report from week one is easy. Reconstructing the same picture in week twelve, after only saving the ones that seemed to matter at the time, is much harder.