Test Security: A Multivariate Outlier Detection Study

• In Event: Test Security and Test Taking Behavior

Abstract

The COVID-19 pandemic disrupted traditional in-person testing models, leading to the rise of remote proctoring. This study utilized the isolation forest algorithm to analyze testing data from a medical licensure exam, aiming to identify both outlier items and candidates. Preliminary results indicate low percentages of anomaly items and candidates.

Authors

• Cecilia Alves, MEDICAL COUNCIL OF CANADA

• Qi Guo, MEDICAL COUNCIL OF CANADA

• Andrea Julie Gotzmann, MEDICAL COUNCIL OF CANADA