Predicting the Effect of Violent Gameplaying to Violent Behavior Intention among Females: Indonesia

This study evaluates the predictive relationship between exposure to violent video games and the intention to exhibit aggressive behavior among female players. By applying machine learning techniques to a secondary dataset of 91 female participants, the research aims to quantify how specific game-playing constructs—such as the intensity and duration of violent interactions within a game—map to hostile and instrumental aggressive intentions. The analysis serves as a tool for parents, educators, and game developers to better understand and mitigate potential negative behavioral outcomes associated with gaming.

The methodology utilizes two primary machine learning models: Decision Tree Regression (DTR) and Adaptive Boosting Tree Regression (AB-DTR). The dataset, sourced from Iowa State University, tracks participants who played either a violent or non-violent game across 25 trials. The models were trained to predict two specific dependent variables: hostile intent (the desire to harm an opponent) and instrumental intent (the desire to control or interfere with an opponent's performance). Performance was evaluated using Root Mean Square Error (RMSE) to determine the accuracy of the models in mapping game-playing variables to behavioral intentions.

The findings demonstrate that both DTR and AB-DTR models can predict aggressive behavior intentions with high accuracy. For the hostile variable, the DTR model achieved an RMSE of 0.0 for both training and testing, while the AB-DTR model showed an RMSE of 0.08 for training and 1.08 for testing. For the instrumental variable, the DTR model recorded an RMSE of 0.0 for training and 2.18 for testing, compared to 0.0 and 3.30 for the AB-DTR model, respectively. These results suggest that machine learning is a viable approach for modeling the short-term effects of violent media on behavioral intentions, providing a foundation for future research into broader player demographics.