Predicting Whistleblowing Intention Using Supervised Machine Learning: Integrating TPB and IEDM in State-Owned Enterprises

Abstract

Whistleblowing plays a critical role in detecting organizational misconduct; however, understanding the determinants of whistleblowing intention remains a challenge. Prior studies predominantly rely on regression or structural equation modeling, which focus on explanatory relationships rather than predictive evaluation. This study addresses this limitation by integrating the Theory of Planned Behavior and the Integrated Ethical Decision-Making Model within a supervised machine learning framework. Data were collected from 382 permanent employees of Indonesian state-owned enterprises (BUMN) using a structured questionnaire. Three classification algorithms—Logistic Regression, Support Vector Machine (SVM), and Random Forest—were implemented to evaluate predictive performance. The results indicate that Random Forest achieved the highest predictive accuracy and discrimination capability. Feature importance analysis reveals that perceived behavioral control is the strongest predictor of whistleblowing intention, followed by ethical awareness and attitude, while subjective norms show comparatively weaker influence. These findings refine TPB by demonstrating the dominant role of perceived behavioral control in high-risk ethical decisions and reinforce the importance of ethical awareness as a cognitive trigger within the IEDM framework. The study contributes by bridging behavioral theory and predictive analytics while offering governance insights for strengthening whistleblowing systems in state-owned enterprises.