Peer-reviewed article
Cyber-pornography use among a sample of Chinese university students: utilizing machine learning and network analysis
Li Li, Zhimin Niu, Junting Hu, Songli Mei, Mark Damian Griffiths ·
Summary
Written by AI, not yet AICS's view · claude-sonnet-5-5This cross-sectional survey looked at cyber-pornography use among Chinese university students in different provinces. It used machine learning and network analysis to find which psychological, family, school and demographic factors are linked to use. Only a small share of students were classed as high-risk users. The authors suggest that ecological systems theory is useful for understanding these patterns and that prevention work should look at how the factors interact.
Key points
- The survey included 1322 university students from different provinces of China (the results text refers to 1332 participants).
- 25 students (1.9%) were classed as high-risk cyber-pornography users.
- The XGBoost model identified 15 important features, including social ostracism, moral disengagement, gender, teacher fairness, self-control, depression, parenting style and loneliness.
- In the network analysis, social ostracism was the strongest node in the total sample and among males, while depression was the strongest node in females.
- The authors conclude that psychological, family, school and demographic factors may be directly or indirectly associated with cyber-pornography use.
Limitations
The cross-sectional design cannot show cause and effect or the direction of the links between factors. The sample is described only as students from different provinces, and the high-risk group is very small (25 students). The participants are university students, so the findings say little about younger children. The document also gives two different sample sizes (1322 and 1332).
Abstract
BACKGROUND: Due to the popularity and convenience of the internet, cyber-pornography use among youth has attracted the widespread attention of scholars. Cyber-pornography use among Chinese university students remains rarely investigated, especially based on ecological systems theory using machine learning and network analysis approaches. Therefore, the present study examined cyber-pornography use through machine learning and network analysis based on ecological systems theory among a sample of Chinese university students. METHODS: A sample of 1322 university students in different provinces of China were surveyed in a cross-sectional study. The survey included cyber-pornography use variables, demographic information, psychological variables, family variables, and school factors. RESULTS: Of 1332 participants, 25 students were classed as high-risk cyber-pornography users (1.9%). The XGBoost model of machine learning identified 15 important features of cyber-pornography use including social ostracism, moral disengagement, gender, teacher fairness, year of study, age, autonomy, self-control, depression, father and mother's parenting style, relatedness, experience of being left behind, loneliness, and family closeness. In the network analysis, social ostracism was the strongest node in the total sample and among males, whereas depression was the strongest node in females. CONCLUSIONS: Psychological, family, school, and demographic factors (e.g. gender) may directly or indirectly be associated with cyber-pornography use among Chinese university students. The results suggest that ecological systems theory, including the microsystem, mesosystem, macrosystem, and chronosystem, may be considered as one of the core theories in understanding cyber-pornography use. In addition, the interaction of related factors also needs to be examined deeply to aid prevention and intervention of cyber-pornography use in the future.