KMAN Publication Inc. (KMANPUB) Journal of Adolescent and Youth Psychological Studies (JAYPS) 2981-2526 7 Serial Number 44 2026 04 10 Predicting Cyberbullying Perpetration from Moral Disengagement, Online Disinhibition, Trait Aggression, and Social Network Density Using Random Forests Predicting Cyberbullying Perpetration from Moral Disengagement, Online Disinhibition, Trait Aggression, and Social Network Density Using Random Forests 1 10 10.61838/kman.jayps.5227 EN Journal Article 2025 10 16 <table> <tbody> <tr> <td> <p><strong>Objective:</strong> The aim of this study was to predict cyberbullying perpetration from moral disengagement, online disinhibition, trait aggression, and social network density using a Random Forest machine learning algorithm among a sample of South African adolescents and young adults.</p> <p><strong>Methods and Materials:</strong> The study employed a quantitative, cross-sectional predictive research design with a sample of participants from three South African provinces. Data were collected using the Cyberbullying Offending Scale, Cyberbullying Moral Disengagement Scale, Online Disinhibition Scale, Buss-Perry Aggression Questionnaire, and a structural proxy questionnaire for social network density. A Random Forest model was trained on of the data and tested on the remaining , utilizing -fold cross-validation for hyperparameter tuning.</p> <p><strong>Findings:</strong> The Random Forest model demonstrated high predictive performance on the testing set ( ), achieving an Accuracy of , Precision of , Recall of , F1-score of , and . Variable importance metrics revealed that trait aggression was the strongest predictor of cyberbullying perpetration (Mean Decrease Gini ), followed by moral disengagement (Mean Decrease Gini ) and online disinhibition. Social network density was the weakest predictor in the model (Mean Decrease Gini ).</p> <p><strong>Conclusion:</strong> Individual psychological factors, specifically trait aggression and moral disengagement, overwhelmingly drive cyberbullying perpetration compared to digital structural environments, emphasizing the need for interventions focused on anger management and cognitive restructuring.</p> </td> </tr> </tbody> </table> https://journals.kmanpub.com/index.php/jayps/article/download/5227/9498