Detecting Illicit Collaboration Using Machine Learning

• In Event: Test Security and Test Taking Behavior

Abstract

This study proposes a framework that combines IRT modeling and machine learning for detecting illicit group collaboration among test takers. In preliminary empirical and simulation studies, this approach achieved high detection rate, which shows promise for efficient pre-screening in educational assessments.

Authors

• Danqi Zhu, FORDHAM UNIVERSITY

• Joseph Grochowalski, COLLEGE BOARD

• Amy H. Hendrickson, COLLEGE BOARD