Association Between Self-Control and Internet Addiction in University Students: A Simulation-Based Network Analysis

Deficits in self-control constitute a core risk factor for internet addiction among university students. To better characterize heterogeneity in this association, the present study adopted a person-centered approach to identify latent self-control profiles and to examine whether internet addiction symptom networks and model-based candidate leverage points differed across profiles. In this cross-sectional survey, 1704 university students completed self-report measures of demographic characteristics, self-control, and internet addiction symptoms. Latent profile analysis (LPA), Ising network analysis, Gaussian graphical model (GGM) analysis, and simulation-based threshold-perturbation modeling using NodeIdentifyR were conducted. These simulated perturbations represented hypothetical model-based scenarios rather than empirical clinical or behavioral interventions. Latent profile analysis identified three self-control subgroups—low, moderate, and high (23.94%, 62.79%, and 13.26%, respectively)—with significant differences in internet addiction severity across groups. In the overall network, CIAS24 emerged as a prominent node, whereas profile-specific differences were observed across self-control subgroups. Model-based simulations suggested that tolerance-related and self-regulation-related nodes may function as distinct leverage points across profiles. This cross-sectional, model-based study suggests that internet addiction symptom organization varies across self-control profiles. The simulation findings should be interpreted as hypothetical and hypothesis-generating rather than as evidence of real-world intervention efficacy.