The test grader matches or exceeds human inter-rater reliability in studies of rubric-based scoring. The system applies rubric criteria consistently and thoroughly, but you review and approve every grade before finalizing.
Accuracy improves when rubrics are clear and specific. Vague criteria like ‘good writing’ are harder for the system (and humans) to score consistently. Detailed rubrics yield the best results.




