Grade a select amount of assignments to create a baseline for our powerful AI. Allow students to mark the remainder of assignments. Kritik will allot each student a Kritik Score based on the accuracy of their grading and automatically tailor the grades they gave to other students based upon their Kritik score.
Calibration & anonymous peer review removes any threat of personal bias or favouritism. This ensures each assignment a student submits is carefully assessed & graded fairly.
The impact of each grader on the final grade varies proportionally, based on their previous performance in grading. Kritik uses advanced data analysis to determine who is a good grader. "TA-level" graders will have a higher weight towards the creation score of peers, increasing the accuracy of the final scores.
Our algorithm uses AI to flag the graders who may have not been accurate in their review and brings them to your attention. Also, we use data to create groups of evaluators that changes activity to activity so that the evaluators across all assignments are homogeneous in their grading ability. By doing this, we ensure that strong and weak graders are evenly distributed among assignments.
