A study on detecting faking response using machine learning in a self-report psychological test

This study investigated methods for detecting faking responses in self-report psychological tests using machine learning techniques. Data were collected under both honest and deceptive conditions, and three algorithms—random forest, support vector machine, and artificial neural network—were applied for prediction. The results indicated that machine learning techniques, particularly artificial neural networks, were effective, achieving more accurate results compared to traditional methods using specific cutoff thresholds. These findings suggest that the choice of detection method may vary depending on the context of the psychological assessment. Furthermore, the study recommends combining machine learning and traditional approaches to enhance the detection of faking responses in psychological tests.