关键词: 青少年，网络游戏成瘾，风险因素和保护因素，增长混合模型，网络分析

Abstract: The problem of adolescent Internet Gaming Disorder (IGD) can be more thoroughly and accurately analyzed from a perspective that combines dynamic, heterogeneous, and holistic elements. This study meets this demand by integrating growth mixture modeling and network analysis. First, to explore the dynamics and heterogeneity of adolescent IGD development, growth mixture modeling was employed to identify subgroups of developmental trajectories in IGD. Second, to address the need for a holistic and heterogeneous analysis of the complex relationships between adolescent IGD and various risk and protective factors, network analysis was used to generate networks of IGD and its multi-domain relationships at different time points for each subgroup. Comparisons were made between the networks and network attributes of the different subgroups at each time point. Third, recognizing that network models may change dynamically over time, network comparison methods were used to analyze the temporal changes in network models and network attributes for each subgroup and compare the changes across subgroups. Lastly, by integrating the results of these comparisons, the study explored the heterogeneous development trajectories, underlying reasons, and developmental characteristics of adolescent IGD. Using a cluster sampling method, a total of 1,852 adolescents from Shandong Province, China, were followed for one year across three assessment waves. Participants were recruited from three junior high schools (Grades 7–9), two general high schools (Grades 10–11), and two vocational high schools (Grades 10–11). Participants completed questionnaires assessing 15 risk and protective factors from family, school, community, and individual domains, as well as an IGD scale. The study performed preliminary analyses using SPSS 26.0, and missing values were imputed using the Expectation-Maximization algorithm. A growth mixture model was constructed using Mplus 8.3. Network analysis and network comparison analysis were performed using R software (version 4.3.1). Results revealed two main IGD developmental trajectory subgroups: the normal development group (characterized by low IGD scores that do not significantly change over time) and the addiction development group (characterized by initially high IGD scores that significantly increase over time). In the early stages of IGD development, no significant differences were found between the two subgroups in terms of overall network connectivity or global strength. Key factors (proximal factors and high centrality nodes) did not differ noticeably, with deviant peer affiliation, peer victimization, parental psychological control, teacher support, community resources, and self-esteem being significant proximal factors. School belonging and self-esteem were stable high-centrality nodes in both groups' networks, and no unique factors were found to affect only the addiction development group. The main distinction between the two subgroups was that the addiction development group had significantly lower scores on all protective factors and significantly higher scores on risk factors compared to the normal development group, suggesting that the abnormal development of adolescent IGD is more likely driven by the accumulation of multi-domain adverse factors in a "low protection-high risks" pattern. In the later stages of IGD development, the overall network connectivity in the addiction development group decreased, and all direct connections between factors and IGD disappeared, indicating a possible hysteresis phenomenon in IGD development. This suggests that, under the cumulative influence of various factors, IGD may enter a self-sustaining equilibrium state. Even when external risks diminish, IGD may not immediately alleviate. This study provides valuable insights into the development trajectories of adolescent IGD, the reasons behind the early rapid development of IGD, and the unique characteristics of the addiction development group in the later stages.

Key words: adolescents, Internet gaming disorder, risk factors and protective factors, growth mixture modeling, network analysis