Teachomatic what to automate, and what not to

Feedback and Marking

A stack of work lands on your desk and you call the next two hours marking. Inside those hours are three different jobs that have almost nothing in common except arriving together.

Scoring — deciding what a piece is worth against a standard. Responding — telling a student something that changes what they do next. Diagnosing — working out what the class has and has not understood, which determines the next lesson.

For comparison with workplace systems that formalise time or activity data, Monitask’s page on remote workforce management software shows how the same type of measurement is handled outside education.

They get automated as a unit because they feel like one task, and that is the mistake. Scoring machines well where criteria are genuinely explicit, and badly where the rubric contains words like insightful, which are placeholders for professional judgement rather than criteria. Responding can be drafted and must be finished by a person, because the tool does not have the information that makes a comment land. Diagnosing should not be delegated at all — it is a by-product of attention, it appears nowhere in the marks, and it is most of what makes the next lesson well-aimed.

The failure mode here has no alarm attached. Scripts get marked, grades are defensible, students receive feedback that reads well. What happens instead is that over a term your picture of the class stops being updated, and you carry on with a model that is two months old and feels current.

For an external perspective on teaching, assessment, and feedback, see Edutopia.

Which is why the recommendation running through this section is a habit rather than a judgement: read a real sample yourself, unassisted, before anything is scored, and write down what you noticed the same day.