The present disclosure relates to a method for monitoring student behavior during remote examinations. The method involves receiving a data stream from a student's device, which may include video, audio, and desktop activity data. By analyzing this data stream, the system detects student behavior events and determines if any constitute violations based on pre-established conditions. When a violation is identified, a unique weight is assigned according to the detection mode, ensuring that the same violation receives different weights depending on how it was detected. The system aggregates weights from multiple violations to calculate a score for the student. Based on this score, an alert is transmitted to notify relevant parties. This approach enables nuanced assessment and real-time alerting of student conduct during remote testing environments.