A mark tells a learner where they are. An error log tells them why — and, more importantly, tells you what to teach next week.
Why error logs fail
Most error logs are lists: "article misuse ×3, subject-verb agreement ×2." They die after two weeks because nobody can turn them into a lesson. The fix is to tag every error with the band descriptor it attacks.
Four buckets, not forty categories
The writing criteria give you the taxonomy for free:
- Task Response — off-topic, under-developed ideas, missing position.
- Coherence & Cohesion — paragraph structure, linking, progression.
- Lexical Resource — word choice, collocation, repetition.
- Grammatical Range & Accuracy — sentence variety, errors per sentence.
Tag the error, not the learner's mood. "Overuses moreover" is Lexical Resource. "Five run-ons in paragraph two" is Grammar.
Making the trend visible
- Record errors per essay against the four criteria.
- At the end of the month, count per learner — not per essay.
- Teach to the mode: the single most frequent error across the class becomes Monday's mini-lesson.
When a learner sees "run-ons: 5 → 3 → 1" across three essays, the progress is self-motivating. When you see the same three errors across the whole class, you know the lesson to plan.
That is the difference between marking essays and managing learning.