When it comes to factors that are most associated with pornography use during adolescence, the most predictive variable is gender, followed by problematic internet use, cannabis and alcohol consumption, family conflict and engaging in certain risky sexual behaviours.
This is according to a study published in the journal Computation, and which was based on the 2023 ESTUDES survey and worked with a sample of 33,543 adolescents. The aim was to determine which personal, family, digital and behavioral variables are most useful for predicting whether an adolescent reports having viewed or used pornography in the previous 30 days.
Detecting patterns with artificial intelligence
The study combined the conventional statistical technique of logistic regression, three machine learning models (XGBoost, LightGBM and CatBoost) and explainable artificial intelligence techniques. These tools are able to detect complex relationships between variables and reveal the weight that each variable carries in the predictions.
The different models produced similar results. The machine learning systems slightly improved some indicators, but the differences compared with logistic regression were small. The authors, therefore, noted that more complex models should not automatically replace traditional statistical techniques. “Artificial intelligence helps us see that this behavior does not appear in isolation, but within a broader profile of digital, family and behavioral factors,” the authors affirmed.
Sex and problematic internet use: the most influential factors
The sex of the individual is, by far, the variable that carries the most weight in the predictions, with boys being more likely to use pornography. Age is also associated with it, but has a lower predictive importance.
After gender, the next most significant factor is problematic Internet use, which includes behaviors such as having difficulty staying offline, sleeping less to be online, neglecting obligations or using the Internet to escape from worries or negative emotions. The researchers pointed out that problematic internet use does not equate to a clinical diagnosis of Internet addiction.
The next most predictive variables are cannabis and alcohol use and family conflict. Some risky sexual behaviors, such as having sex without a condom or having a sexual encounter that the teenager regrets the next day, are also slightly predictive. In contrast, parental control has a much more limited effect when all the variables are analyzed together.
The study also identified differences between boys and girls. Among boys, alcohol consumption is the most important predictor once the individual’s sex is excluded from the analysis, whereas among girls it is problematic internet use. The weight of family conflict also varies. Age-related differences, however, are smaller.
More comprehensive prevention
The research team insisted that although the results show associations and predictive patterns, cause-and-effect relationships cannot be established. For this reason, the models should not be used to diagnose, label or sanction specific adolescents, but rather to better understand the phenomenon and to guide educational and preventative policies.
The conclusions pointed towards programs that combine sex and relationship education, digital literacy, critical reflection on pornographic content, substance misuse prevention, the promotion of safe sexual practices and the strengthening of family communication.
“Pornography should not be addressed as a phenomenon disconnected from the other dimensions of adolescent life. It is part of a broader ecosystem of digital practices, family relationships, substance use and sex education,” the authors summarized.
The study was carried out by the researchers Jorge de Andrés-Sánchez, from the URV’s Department of Business Management, and Ángel Belzunegui-Eraso, Inma Pastor-Gosálbez and Anna Sánchez-Aragón, from the URV’s Department of History and Art History, all members of the Social and Business Research Laboratory.
