Unravelling social network usage patterns: A study based on unsupervised learning

We are facing a growing concern regarding the consequences of Problematic Social Network Use (PSNU). Therefore, the aim of the present study was to explore the profiles of social network (SN) users and contrast their differences across variables linked to PSNU in the literature. A sample of 726 participants aged 16 and above (77.0% female) residing in Spain was analysed using Latent Class Analysis (LCA). ANOVA and chi-square test were employed to investigate differences in sociodemographic variables, digital preferences, impulsivity, emotional intelligence, empathy and aggression among the latent classes. The three-class model, which proved the most parsimonious, identified functional, risky, and problematic users. Notably, greater differences were observed between functional and risky users regarding impulsivity and emotional regulation, whereas differences in empathy and aggression were more pronounced between problematic users and the other two groups. The study provides relevant information about the characteristics of different groups of SN users. This information may be useful for the early detection of inappropriate online behaviours that may lead to PSNU, as well as for identifying users who may have already developed it.