Identifying risk profiles and systemic architecture of online aggressive behavior among college students: a large-scale cross-sectional study.

Identifying risk profiles and systemic architecture of online aggressive behavior among college students: a large-scale cross-sectional study - PMC Skip to main content An official website of the United States government Here's how you know Here's how you know Search Log in Primary site navigation Search Log in Search PMC Full-Text Archive Search in PMC Journal List User Guide As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: PMC Disclaimer | PMC Copyright Notice BMC Public Health . 2026 Apr 21;26:1795. doi: 10.1186/s12889-026-27477-z Identifying risk profiles and systemic architecture of online aggressive behavior among college students: a large-scale cross-sectional study Lijun Zhu 1, # , Yan Chen 1, # , Mengyun Huang 2, # , Zhengmei Fang 1 , Yurui Zhou 1 , Jiajing Zhao 1 , Yichen Li 1 , Yejun Xie 1 , Long Hua 1 , Liying Wen 1 , Weiwei Chang 1 , Yuelong Jin 1, ✉ , Yingshui Yao 1, ✉ Author information Article notes Copyright and License information 1 Department of Epidemiology and Biostatistics, School of Public Health, Wannan Medical College, Institute of Chronic Disease Prevention and Control, No. 22, Wenchang Road, Wuhu, 241002 China 2 Department of Clinical Medicine, Anhui College of Traditional Chinese Medicine, Wuhu, 241003 China ✉ Corresponding author. # Contributed equally. Received 2026 Feb 19; Accepted 2026 Apr 16; Collection date 2026. © The Author(s) 2026 Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ . PMC Copyright notice PMCID: PMC13235130  PMID: 42015144 Abstract Background Online aggressive behavior (OAB) is a growing public health concern among young adults. However, most studies focus on isolated risk factors rather than their complex interplay. This study aimed to identify hierarchical risk profiles and map the systemic architecture of OAB among Chinese college students. Methods A cross-sectional survey was conducted with 5,431 college students in China. We integrated Classification Tree Analysis (CTA) using the CHAID algorithm to identify high-risk subgroups and Network Analysis (Ising model) to map the conditional dependencies among 15 OAB items and 7 key risk factors. Results The overall prevalence of OAB was 56.8%. CTA identified video game engagement as the primary stratifying variable ( χ² = 280.24, p < 0.001), revealing a clear risk gradient: competitive gaming (64.5%) > non-competitive gaming (60.5%) > no gaming (35.7%). Alcohol use and internet use further stratified these groups. The highest-risk profile (89.0% prevalence) was identified among non-competitive gamers who used alcohol and reported poor academic performance. The network analysis (95 edges; density = 0.411) identified competitive gaming as the central hub (highest Strength) and “Hacking to steal identities” as a highly influential node. “Insulting others in games” emerged as the critical bridge node (highest Bridge Strength). All network indices showed high stability (CS-coefficient ≥ 0.75). Conclusions OAB is highly prevalent among Chinese college students and its risk is structured hierarchically, with competitive gaming acting as a key high-risk context. Public health strategies could benefit from moving beyond universal approaches to targeted interventions aimed at specific high-risk profiles and disrupting central nodes and bridging behaviors within the OAB system. Supplementary Information The online version contains supplementary material available at 10.1186/s12889-026-27477-z. Keywords: Online aggression, Classification tree analysis, Network analysis, College students, Risk factors Introduction Online aggressive behavior (OAB) has become a significant public health concern in our digital society, posing particular risks to college students who are highly active online [ 1 , 2 ]. While many risk factors have been identified, existing research is often limited by methods that struggle to capture the complex, interacting nature of these risks. This study addresses this gap by integrating two advanced methods: Classification Tree Analysis (CTA) and Network Analysis (NA). By moving beyond traditional linear models [ 3 , 4 ], we aim to map the hierarchical relationships and interconnected structure of OAB risk factors among Chinese college students. The goal is to provide a foundation for developing more precise prevention and screening strategies. The pervasiveness and theoretical context of online aggression Global digitalization has made online spaces essential for connection but has also increased the risk of OAB [ 5 , 6 ]. OAB includes a range of harmful online behaviors, such as cyberbullying and defamation, which are strongly associated with psychological distress, including anxiety and depression [ 7 ]. Theoretically, these behaviors are grounded in the General Aggression Model (GAM), which posits that aggression is an outcome of the complex interaction between person (e.g., traits, internal states) and situation factors (e.g., environmental cues) [ 8 ]. The Chinese context provides a critical setting for examining this model. As of January 2025, China’s internet user base reached 1.092 billion, with a penetration rate of 77.5% [ 9 ]. Chinese college students are a particularly high-risk group due to their extensive engagement with social media [ 10 ]. This environment, often characterized by anonymity and extended screen time, may contribute to a “disinhibition effect” that alters normal social controls [ 11 ]. Empirical evidence highlights the urgency of this issue, with prevalence rates of cyber aggression among young adults reported to be between 27.5% and 70% globally [ 12 – 14 ]. Victims often suffer from lasting psychological impacts [ 15 , 16 ]. Given this context, it is crucial to understand the mechanisms underlying the co-occurrence of OAB and its risk factors. To systematically investigate these mechanisms, the selection of risk factors in this study was strictly guided by the GAM framework [ 8 ]. Specifically, person factors included sociodemographic variables (e.g., gender) and familial contexts, such as parent-child relationships, as strained family dynamics are established correlates of aggressive behavior [ 17 ]. For situation factors, we focused on contexts prevalent among college students. These included high-risk digital environments like competitive gaming, which is strongly linked to aggression [ 18 ], and physiological disinhibitors like alcohol consumption, known to impair impulse control and be associated with aggression [ 19 ]. By operationalizing these theory-driven variables, our study aims to map the specific interactive pathways through which these factors converge to shape OAB risk. Methodological gaps: from linear associations to systemic complexity Current research on OAB determinants predominantly uses traditional regression models [ 20 , 21 ]. While useful for identifying individual predictors, these methods have key limitations. First, regression models typically assume that risk factors work independently and additively. However, the GAM suggests that aggression is often linked to synergistic interactions (e.g., a specific personality trait combined with a specific situational stressor) [ 22 ]. Second, linear models often focus on the degree of a behavior. In contrast, from a public health perspective, the priority is often to distinguish “at-risk” from “low-risk” individuals for efficient, large-scale screening [ 23 , 24 ]. To better understand OAB, we need methods that can model these complex interactions and identify distinct risk groups. An integrated systems approach: CTA and NA To overcome these limitations, our study integrates two complementary methods: Classification Tree Analysis (CTA) and Network Analysis (NA). This approach aligns with the theoretical complexity of the GAM. CTA is a powerful tool for identifying hierarchical structures. As a non-parametric method, it excels at modeling complex, non-linear interactions without the restrictive assumptions of linear regression [ 25 ]. CTA works by recursively partitioning the data to identify specific combinations of person and situation factors that best distinguish high-risk individuals from low-risk ones, thus creating clear “risk profiles” [ 26 , 27 ]. It helps answer the question: Who is most at risk? However, identifying risk profiles is not enough; we also need to understand how risk factors and behaviors are interconnected. NA addresses this by conceptualizing psychological phenomena as a system of interconnected nodes (e.g., specific behaviors, risk factors) [ 28 , 29 ]. NA allows us to visualize this structure, identifying central factors (hubs) and key pathways that may be crucial for maintaining the system [ 30 , 31 ]. It helps answer the question: How are these risk factors and behaviors structurally related? The present study Building upon this integrated framework (see Fig. 1 ), this study uses a dual-analytic approach to investigate the risk architecture of OAB among Chinese college students. First, we use Classification Tree Analysis (CTA) to delineate hierarchical risk profiles, which we define as the multi-level categorization of high-risk subgroups based on their unique combination of interacting predictors. Subsequently, we employ Network Analysis (NA) to map the underlying systemic architecture, conceptualized as the complex network of statistical dependencies among specific aggressive behaviors and their associated risk factors. Fig. 1. Open in a new tab Integrated systems approach to online aggressive behavior (OAB) risk architecture. Note . GAM = General Aggression Model; CTA = Classification Tree Analysis; NA = Network Analysis; OAB = Online Aggressive Behavior This integrated methodology allows us to: Identify hierarchical risk profiles for OAB to provide a tool for efficient population screening. Map the network structure to identify central risk factors and potential targets for intervention. Offer a data-driven framework to inform the development of precision prevention strategies targeting both high-risk subgroups and key systemic factors. Methods Procedure and participants This study employed a cross-sectional design. Data collection occurred between September 2022 and September 2023 at four higher education institutions in Anhui Province, China, comprising two vocational colleges and two undergraduate universities. A stratified cluster sampling method was used, with academic discipline and year of study serving as strata. The target sample size was determined through an a priori power analysis based on a previously reported methodology [ 32 ]. The calculation was based on the formula , with a significance level (α) of 0.05 (U α/2 = 1.96), a maximum allowable error (δ) of 0.1P, and a conservative estimated prevalence of aggressive behavior (P) of 13% from prior literature. These parameters yielded an initial required sample size of 2,600. To account for the design effect of cluster sampling (adjusted by 50%) and potential invalid responses (10%–20%), the target sample size was expanded to 5,700 to ensure robust statistical power. Data were collected via in-person, paper-and-pencil surveys administered in classroom settings. This standardized procedure, overseen by trained research assistants, was designed to maximize data quality by minimizing distractions and ensuring consistent instructions. The survey required approximately 20–25 min for completion. A total of 5,700 questionnaires were distributed; however, 269 were excluded due to incomplete responses (missing data) or evident invalidity, such as obvious patterned responding (e.g., selecting the same option for all items). We employed a complete-case analysis approach, utilizing only datasets with entirely complete and logically consistent responses. Ultimately, 5,431 valid cases were retained for the final analysis, yielding a response rate of 95.28%. The study protocol was approved by the Institutional Review Board of Wannan Medical College (Approval No. 2022AH051208). All participants provided written informed consent after being fully briefed on the study’s objectives and their right to withdraw at any time. Measures Online aggressive behavior (OAB) OAB was assessed using the 15-item Online Aggressive Behavior Scale (OABS) [ 33 ], which has been validated for use in Chinese populations [ 34 ]. The scale measures two dimensions: Overt Aggression (7 items; e.g., “Posted threatening messages”) and Relational Aggression (8 items; e.g., “Spread rumors online to harm someone’s reputation”). Participants rated the frequency of these behaviors over the past six months on a 4-point scale (1 = Never to 4 = Always). In the current sample, the OABS demonstrated excellent reliability and validity. The internal consistency was high, with a Cronbach’s α of 0.921 for the full scale (0.816 for Overt and 0.890 for Relational). Regarding construct validity, the Kaiser-Meyer-Olkin (KMO) measure was 0.966 for the total scale (0.887 for Overt and 0.927 for Relational), and the Bartlett’s test of sphericity was significant ( p < 0.001), indicating that the data were highly suitable for factor analysis and the scale structure was robust in the target population. Two distinct scoring procedures were applied to the OABS data for our analyses. First, for the Classification Tree Analysis, a mean score was calculated for each participant. Following established precedents [ 35 , 36 ], participants with a mean score > 1.0 were classified as “OAB-positive.” As the scale anchor is 1 (Never), this sensitive cut-off identifies any individual who has engaged in at least one aggressive act, making it suitable for broad-based screening from a primary prevention standpoint. Second, for the network analysis, each of the 15 OABS items was dichotomized to meet the requirements of the Ising model. Responses of “1” (Never) were recoded as 0 (absence), while responses from “2” (Seldom) to “4” (Always) were recoded as 1 (presence). This dichotomization strategy is consistent with recent network analysis studies using this scale [ 37 ] and is theoretically grounded in capturing the critical distinction between non-engagement versus any engagement in aggressive behavior. While this approach reduces variance, it is a prerequisite for the Ising model, which is specifically designed to estimate networks of pairwise interactions from binary data [ 38 , 39 ]. This binary coding provides a clear and interpretable network structure focused on the co-occurrence of aggressive acts. Sociodemographic and behavioral measures In line with the General Aggression Model (GAM) framework, we collected data on key person and situation factors. All sociodemographic and behavioral characteristics were assessed via a structured self-report questionnaire. Sociodemographic variables included gender, residence (urban, county/town, rural), only-child status (yes/no), family structure (nuclear/non-nuclear), and parental education levels (categorized as ≤9th grade or ≥High school). Self-reported academic performance was rated as upper, middle, or lower. Behavioral factors were measured using single-item self-reports. Smoking and alcohol use were recorded dichotomously (yes/no). Internet use and physical exercise duration were categorized into three levels. Video game engagement was classified as none, competitive gaming, or non-competitive gaming. The perceived quality of parent-child relationships was rated as good, fair, or poor. See Supplementary Appendix 1 for the specific wording of these single-item questions. Data analysis Our data analysis was conducted in three phases. Preliminary analyses were performed in SPSS 26.0, while CTA and NA were conducted in R (version 4.2.1). All tests were two-tailed, with a significance level of α = 0.05. Phase 1: Preliminary Analysis. Descriptive statistics were calculated for all variables. Associations between the binary OAB status and categorical predictors were examined using Pearson’s chi-square ( χ² ) tests. Phase 2: Classification Tree Analysis (CTA). We used the Chi-Squared Automatic Interaction Detector (CHAID) algorithm to identify hierarchical predictors of OAB. CHAID was selected for its robustness with categorical predictors and its interpretable, multi-way splits. To prevent overfitting and ensure model stability, we implemented a 10-fold cross-validation procedure [ 40 ]. The model’s growth was constrained by pre-specified stopping rules: a maximum tree depth of three levels, a minimum of 100 cases required in a parent node for a split, and a minimum of 50 cases in any resulting child node. A Bonferroni correction was applied to adjust p-values at each split. Phase 3: Network Analysis (NA). To ensure methodological continuity and parsimony, the network analysis included only the 15 OAB items and the significant predictors identified in the preceding CTA. Given the binary data, we estimated an Ising model using the IsingFit R package [ 38 ] to compute a network of regularized partial correlations. In this network, nodes represent variables, and edges represent their statistical relationships after controlling for all other nodes. Two centrality indices were calculated to identify influential nodes: Strength, the sum of the absolute weights of a node’s connections, indicating its overall influence; and Expected Influence (EI), the sum of all edge weights (positive and negative), estimating its net activating or deactivating effect on the system. The robustness of centrality estimates was assessed using a case-dropping bootstrap procedure ( n = 1,000) to compute the Correlation Stability (CS) coefficient [ 41 ], with values > 0.25 considered acceptable and > 0.50 stable. Finally, we calculated each node’s predictability—the proportion of its variance explained by its neighbors, analogous to R² in regression [ 42 ]. Results Sample characteristics and bivariate associations Among the 5,431 participants, 56.8% ( n = 3,083) were classified as having engaged in online aggressive behavior (OAB), with 48.2% reporting at least one act of overt aggression (OA) and 31.6% reporting relational aggression (RA) in the past six months (Fig. 2 ). As detailed in Table 1 , subsequent bivariate analyses revealed significant associations between OAB status and a range of characteristics. OAB prevalence was significantly higher among male students compared to female students (64.3% vs. 49.2%, χ² = 125.55, p < 0.001). Behavioral factors were particularly strong predictors; OAB was more prevalent among students who smoked (75.9% vs. 55.5%), used alcohol (72.2% vs. 52.8%), reported heavy internet use (66.8%), or engaged in competitive video games (64.5%) (all p < 0.001). Finally, several familial factors, including non-nuclear family structures, lower paternal education, and poorer father-child relationships, were also significantly associated with a higher prevalence of OAB (all p < 0.05). Fig. 2. Open in a new tab Reporting status of online aggressive behavior among college students. Note . OA = Overt Aggression; RA = Relational Aggression; OABS = Online Aggressive Behavior Scale Table 1. Prevalence of online aggressive behavior among college students by sociodemographic, familial, and behavioral characteristics Characteristics N Non-Online Aggression Group (Non-OAG) Online Aggressive Behavior Group (OAB) χ2 P Gender Male 2726 974(35.7) 1752(64.3) 125.554 < 0.001 Female 2705 1374(50.8) 1331(49.2) Residence Urban 1284 567(44.2) 717(55.8) 0.697 0.706 County/Town 1635 697(42.6) 938(57.4) Rural 2512 1084(43.2) 1428(56.8) Academic Performance Upper 1439 739(51.4) 700(48.6) 65.363 < 0.001 Middle 2924 1228(42.0) 1696(58.0) Lower 1068 381(35.7) 687(64.3) Academic Load Light 1444 650(45.0) 794(55.0) 4.548 0.103 Moderate 2767 1158(41.9) 1609(58.1) Heavy 1220 540(44.3) 680(55.7) Personality Introverted 4103 1743(42.5) 2360(57.5) 3.868 0.049 Extroverted 1328 605(45.6) 723(54.4) Only child Yes 1575 611(38.8) 964(61.2) 17.816 < 0.001 No 3856 1737(45.0) 2119(55.0) Family type Nuclear family 3816 1686(44.2) 2130(55.8) 4.710 0.030 Non-nuclear family 1615 662(41.0) 953(59.0) Father’s education ≤ 9th grade 3489 1425(40.8) 2064(59.2) 22.722 < 0.001 ≥High school diploma 1942 923(47.5) 1019(52.5) Mother’s education ≤ 9th grade 3811 1601(42.0) 2210(58.0) 7.791 0.005 ≥High school diploma 1620 747(46.1) 873(53.9) Father-child relationship Good 4339 1965(45.3) 2374(54.7) 38.466 < 0.001 Fair 951 340(35.8) 611(64.2) Poor 141 43(30.5) 98(69.5) Mother-child relationship Good 4748 2085(43.9) 2663(56.1) 7.114 0.029 Fair 524 202(38.5) 322(61.5) Poor 159 61(38.4) 98(61.6) Smoking Yes 344 83(24.1) 261 (75.9) 54.623 < 0.001 No 5087 2265(44.5) 2822(55.5) Alcohol use Yes 1112 309(27.8) 803(72.2) 135.923 < 0.001 No 4319 2023(47.2) 2280(52.8) Exercise time Less 3464 1497(43.2) 1967(56.8) 0.008 0.996 Moderate 1264 546(43.2) 718(56.8) More 703 305(43.4) 398(56.6) Internet use Less 342 203(59.4) 139(40.6) 131.669 < 0.001 Moderate 3251 1534(47.2) 1717(52.8) More 1838 611(33.2) 1227(66.8) Video games 280.241 < 0.001 None 1184 761(64.3) 423(35.7) Competitive 2256 800(35.5) 1456(64.5) Non-competitive 1991 787(39.5) 1204(60.5) Open in a new tab χ 2 represents the chi-square statistic used for comparing categorical variables. Data are presented as n (%) Non-OAG Non-Online Aggression Group, OAB Online Aggressive Behavior Group Hierarchical risk structure: classification tree analysis A CHAID classification tree analysis was conducted to identify hierarchical risk profiles for OAB. The model’s robustness was confirmed via 10-fold cross-validation, which yielded a stable risk estimate of 0.371, closely matching the resubstitution risk of 0.368 and indicating good generalizability. The final model achieved an overall predictive accuracy of 63.2%, demonstrating high sensitivity for identifying OAB-positive students (88.7%) but lower specificity for non-aggressive students (29.8%). The analysis produced a three-level hierarchical model that stratified the sample into 12 terminal nodes (Fig. 3 ). The primary splitting variable was video game engagement ( χ² = 280.24, p < 0.001), establishing it as the most powerful differentiator. This initial split created a clear risk gradient: students playing competitive games had the highest OAB prevalence (64.5%; Node 1), followed by non-competitive gamers (60.5%; Node 2), while non-gamers had the lowest prevalence (35.7%; Node 3). Fig. 3. Open in a new tab Classification tree analysis of factors influencing college students’ online aggressive behavior. Note . Adj. P-value = Adjusted P-value; Chi-square = χ 2 statistic; df = Degrees of freedom. The target variable is Online Aggressive Behavior, where “Yes” indicates the presence of OAB and “No” indicates the absence of OAB At the second level, risk pathways diverged. Among competitive gamers (from Node 1), internet use was the next significant predictor ( χ² = 64.10, p < 0.001); those with “More” internet use showed a heightened OAB prevalence of 75.4% (Node 4). For both non-competitive gamers (from Node 2) and non-gamers (from Node 3), alcohol use emerged as the key stratifying factor ( χ² = 37.35 and χ² = 23.36, respectively; both p < 0.001), consistently elevating OAB risk. At the third level, the model further stratified subgroups based on factors including father-child relationship, academic performance, and internet use, which successfully isolated the profiles with the most extreme risk levels. The highest-risk group (89.0% OAB prevalence) comprised students who played non-competitive video games, used alcohol, and reported lower academic performance (Node 17). Conversely, the lowest-risk profile (21.6% OAB prevalence) was found among non-gamers who did not use alcohol and reported less internet use (Node 20). These pathways highlight the complex interplay of digital habits, substance use, and personal context in shaping OAB risk. Network structure and centrality An Ising model network was estimated to map the relationships among 22 nodes (15 OAB items and 7 significant predictors from the CTA), resulting in 95 non-zero edges and a network density of 0.411 (Fig. 4 ). The network topology revealed a clear community structure, with nodes partitioned into two distinct clusters: OAB items (A1–A15) and influencing factors (B1–B7). Fig. 4. Open in a new tab Network structure of online aggressive behaviors and associated factors among college students. Note . A1–A15 (purple) represent Online Aggressive Behavior items (e.g., A1 = Insult in games). B1–B7 (green) represent influencing factors identified from the CTA (e.g., B1 = Competitive gaming). Edge thickness corresponds to the partial correlation weight. Blue edges indicate positive correlations; red edges indicate negative correlations The strongest inter-community edge connected Competitive gaming (B1) and Insult in games (A1) (weight = 1.81). A strong positive edge also linked Non-competitive gaming (B2) with Insult in games (A1) (weight = 1.43). In contrast, a weaker negative edge was found between Competitive gaming (B1) and Speak badly to friends (A9) (weight = -0.38). Centrality analysis identified nodes with key structural roles (Fig. 5 ). Competitive gaming (B1) had the highest Strength, identifying it as the network’s most influential hub. Hack to steal identities (A8) showed the highest Expected Influence, marking it as a highly influential node within the estimated network structure (Fig. 5 A). Finally, bridge centrality analysis pinpointed Insult in games (A1) as the primary bridge node, serving as the strongest statistical pathway between the influencing factors and the OAB behaviors (Fig. 5 B). Fig. 5. Open in a new tab Centrality and bridge centrality indicators of the online aggression network. Note . A Displays Node Strength and Expected Influence (EI). B Displays Bridge Strength. Centrality indices identify nodes with high structural importance. A1–A15 and B1–B7 correspond to variables in Fig. 4. Values are standardized z-scores The stability of these findings was confirmed via a case-dropping bootstrap procedure (Fig. 6 ). The Correlation Stability (CS) coefficients were excellent for Strength (CS = 0.75), Bridge Strength (CS = 0.75), and edge weights (CS = 0.75), all far exceeding the recommended threshold of 0.50. This indicates that the estimated network structure and the identified central and bridge nodes are highly reliable. Fig. 6. Open in a new tab Stability assessment of the strength and bridge strength centrality measures. Note . Stability analysis of centrality measures using case-dropping bootstrap. The x-axis is the percentage of the sample retained. The y-axis is the correlation between original and subsampled centrality indices. Shaded areas are 95% confidence intervals. The high Correlation Stability (CS) coefficient (CS > 0.75) indicates excellent stability Discussion By integrating classification tree and network analyses, this study offers a multi-layered model of the risk architecture of OAB among Chinese college students. Our findings suggest a hierarchical structure where video game engagement—particularly in competitive contexts—emerges as a primary differentiating factor. This risk is associated with other factors such as alcohol use, internet use time, challenging family dynamics, and academic difficulties. The network analysis further illuminated this structure, highlighting competitive gaming as a central hub and in-game insults as a key bridge connecting this context to a cluster of aggressive acts. This dual-analytic approach moves beyond simple risk-factor lists to a more nuanced, systemic map of the associations underlying OAB. The centrality of gaming as a high-risk context Our classification tree model identified competitive gaming as the most powerful single differentiator of OAB risk, a finding consistent with a growing body of literature [ 43 , 44 ]. For instance, a recent meta-analysis reported a positive association between video gaming and aggression among youth [ 45 ]. First, the high-stakes, frustrating environment of competitive games may foster hostile cognitions and situational aggression [ 46 , 47 ]. Second, individuals with pre-existing aggressive tendencies may self-select into competitive online environments that align with or gratify these traits [ 48 , 49 ]. From a public health perspective, however, the direction of causality is secondary to the empirical finding that competitive gaming functions as a high-risk ecological niche. Whether as a precipitating factor or a magnet for aggression, this context represents a critical “hotspot” where OAB is concentrated, making it a pragmatic and vital point for screening and intervention efforts. The neurobiological evidence suggesting links between problematic gaming and deficits in impulse-control circuits [ 50 , 51 ] further reinforces the view of this population as particularly vulnerable, irrespective of the initial causal mechanism. Interacting risk factors: disinhibition and personal vulnerabilities Our models demonstrate that the risk associated with gaming is not monolithic but is amplified by other factors. Alcohol use emerged as a pivotal secondary node strongly correlated with increased OAB risk across both gaming and non-gaming populations. This supports a disinhibition model where alcohol impairs executive functions by disrupting prefrontal-limbic communication, thereby weakening control over aggressive impulses [ 52 ]. While some research has linked alcohol use to specific online transgressions like privacy invasion [ 53 ], its neurobiological impact suggests a broader effect. Chronic alcohol use can induce atrophy and dopaminergic dysregulation in the same prefrontal-cingulate circuits implicated in impulse control, creating a generalized vulnerability to a range of impulsive online behaviors [ 54 ]. Oxidative stress-mediated damage in these regions further compromises emotional regulation, creating a neurobiological substrate that synergizes with the situational triggers present in online environments to facilitate OAB [ 55 ]. In addition to substance use, the sheer amount of time spent online also appears to be a critical factor. As our tree model showed, higher internet use further stratified risk within the competitive gamer group, a finding consistent with studies by Xiao et al. [ 56 ] and Yamada et al. [ 57 ], which link more time online to increased cyber-aggression and interpersonal conflict. Furthermore, offline personal vulnerabilities played crucial roles in specific contexts. Among competitive gamers with moderate internet use, a poor father-child relationship was associated with a markedly higher prevalence of OAB. This supports extensive research highlighting paternal relationships as a key factor in moderating externalizing behaviors [ 58 , 59 ]. A strained father-child relationship may signify a deficit in emotional support and effective conflict-resolution modeling, potentially impairing an individual’s ability to cope with online frustrations [ 60 , 61 ]. Similarly, poor academic performance was a powerful predictor within a specific high-risk profile (non-competitive gamers who use alcohol). This finding is consistent with literature linking academic difficulties, a potent source of chronic stress and frustration, to aggressive behaviors [ 62 ]. For students struggling academically, the online world may serve as an alternative venue to express frustration or seek a sense of control, which can manifest as aggression [ 63 , 64 ]. These findings underscore that OAB risk is shaped by a complex interplay of online activities, substance use, and offline personal vulnerabilities. From risk profiling to a system-informed prevention framework Our integrated methodology offers complementary implications for public health strategy. First, the classification tree provides a data-driven tool for hierarchical risk profiling. This aligns with a precision public health approach, which seeks to identify more specific sub-populations to deliver tailored interventions, moving beyond one-size-fits-all strategies [ 65 , 66 ]. For example, students in the highest-risk profiles may require more intensive support, whereas those at moderate risk could benefit from targeted digital citizenship education. Second, the network analysis offers a structural map that can generate hypotheses for future intervention research. (1) The Central Hub: The centrality of competitive gaming (B1) suggests it is structurally important. Interventions might not aim for cessation but could test strategies to modify the gaming context, such as promoting cooperative gameplay or platforms with strong pro-social norms.(2) A Potential Amplifier: The high expected influence of Hacking to steal identities (A8) suggests it may function as a behavioral amplifier. As a premeditated act [ 67 ], its occurrence might reinforce an individual’s sense of aggressive efficacy, potentially facilitating other behaviors [ 68 ]. This suggests that targeted technical and legal countermeasures against internet-based identity theft could be a fruitful area for future research to test its broader impact on the OAB system. (3) A Critical Bridge: The identification of In-game insults (A1) as the primary bridge node is critical. It suggests that this behavior is the strongest statistical link between influencing factors (e.g., competitive gaming) and overt aggression [ 69 , 70 ]. This highlights a potential leverage point: focusing on monitoring and de-escalating toxic in-game communication could hypothetically disrupt the strongest pathway connecting risk factors to aggressive acts, a premise that should be tested in future longitudinal designs. An intriguing nuance was the negative edge between competitive gaming (B1) and a form of relational aggression (A9, speaking badly of friends). This may suggest a functional substitution, where individuals in this context favor direct, overt aggression over indirect forms. This underscores the need for tailored, context-aware interventions. Collectively, these network findings provide a rationale for developing and testing multi-pronged strategies that target the OAB system at its most structurally important points. Limitations and future directions This study has several limitations. First, the cross-sectional design precludes causal inferences; our findings reflect statistical associations, not causal pathways [ 41 , 71 ]. Second, the use of single-item self-reports for behavioral predictors may introduce measurement error and social desirability bias. Future research should employ validated, multi-item scales. Third, our sample was drawn from a specific region in China, and the findings require replication in diverse populations to assess generalizability. Finally, methodological choices regarding variable operationalization present trade-offs. Our sensitive threshold for OAB (any occurrence) was chosen to maximize screening utility but resulted in a high prevalence rate. Using a stricter, clinical cutoff would likely yield different risk profiles. Similarly, dichotomizing items for network analysis, while necessary for the model, reduced variance in behavior frequency. Therefore, our models are best understood as mapping the co-occurrence of risk factors and aggressive acts, rather than their severity or escalation. Conclusions Despite these limitations, this study provides an integrated model of the risk architecture of OAB. By combining classification tree and network analyses, we show that OAB is associated with a highly interconnected system of factors, not just a list of isolated risks. This system appears to be centered on the context of competitive gaming, amplified by factors like alcohol use, and shaped by background vulnerabilities. Our findings advocate for a shift in prevention research—from targeting individual factors in isolation to designing and testing strategies that address the interconnected structure of OAB at its most central and influential points. Supplementary Information Supplementary Material 1. (23.7KB, docx) Acknowledgements We are very grateful to all the students who participated in this study for their time and cooperation. Abbreviations CHAID Chi-Squared Automatic Interaction Detector CS Correlation Stability CTA Classification Tree Analysis EI Expected Influence GAM General Aggression Model KMO Kaiser-Meyer-Olkin NA Network Analysis OA Overt Aggression OAB Online Aggressive Behavior OABS Online Aggressive Behavior Scale RA Relational Aggression Authors’ contributions LZ, YC, and MH conceived the study, analyzed the data, and drafted the manuscript. ZF, YZ, JZ, YL, YX, LH, LW, and WC participated in data collection and coordination. YJ and YY designed the study, supervised the research, and revised the manuscript critically for important intellectual content. All authors read and approved the final manuscript. Funding This study was supported by the Scientific Research Project of Anhui Provincial Institutions of Higher Education (Philosophy and Social Sciences) (No. 2022AH051208); Research Topic on Innovative Development of Social Sciences in Anhui Province (No. 2021CX105); Anhui Province Young and Middle-aged Teachers’ Training Action Program for Discipline (Specialty) Leaders Cultivation Project (No. DTR2024031); Wannan Medical College Young and Middle-aged Research Fund Project (No. WK2024ZQNZ17); and Key R&D and Achievement Transformation Projects of Science and Technology Projects in Wuhu City (No. 2023yf103). The funders had no role in study design, data collection, data analysis, data interpretation, or report writing. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate The study was conducted in accordance with the Declaration of Helsinki. Ethical approval was granted by the Ethics Committee of Wannan Medical College (Reference No. LL-2022AH051208). Written informed consent was obtained from all participants included in the study. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Lijun Zhu, Yan Chen and Mengyun Huang contributed equally to this work. Contributor Information Yuelong Jin, Email: jinyl0803@wnmc.edu.cn. Yingshui Yao, Email: yaoyingshui@wnmc.edu.cn. References 1. Ding H, Xu Y. Online aggressive behavior, self-harm behavior, and social anxiety: the mediating effect of social network sites addictive tendency and the moderating effect of sex. Psychiatry Investig. 2025;22(9):1020–30. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 2. Ruan QN, Ni FR, Yu RY, Lin Y. Sex-specific psychological mechanisms of cyber reactive aggression: evidence from Chinese college students. Front Psychol. 2026;17:1764030. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 3. Xia Y, Ma Z. Network structure of the links between extracurricular time-use and delinquent behaviors: Moving forward and beyond linear relations. Child Dev. 2023;94(6):1697–712. [ DOI ] [ PubMed ] [ Google Scholar ] 4. Zeng R, Zhao M, Hu M, Liu X, Chang H. Influencing factors and network analysis of aggressive behavior among Chinese college students based on propensity score matching. Chin J Sch Health. 2025;46(03):358–63. [ Google Scholar ] 5. Guo Y, Cai S, Liang F. The prevalence of cyberbullying and the association between cyberbullying emotional exhaustion and perception of collective efficacy among Chinese college students. J Affect Disord. 2024;362:145–51. [ DOI ] [ PubMed ] [ Google Scholar ] 6. Yen JL, Chamanadjian C. Cyberbullying and online aggression. Pediatr Clin North Am. 2025;72(2):333–49. [ DOI ] [ PubMed ] [ Google Scholar ] 7. Li C, Wang P, Martin-Moratinos M, Bella-Fernandez M, Blasco-Fontecilla H. Traditional bullying and cyberbullying in the digital age and its associated mental health problems in children and adolescents: a meta-analysis. Eur Child Adolesc Psychiatry. 2024;33(9):2895–909. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 8. DeWall CN, Anderson CA, Bushman BJ. The general aggression model: Theoretical extensions to violence. Psychol violence. 2011;1(3):245. [ Google Scholar ] 9. The 55th. Statistical Report on Internet Development in China was released. Media Forum. 2025;9(02):121. [ Google Scholar ] 10. Qin N, Shi S, Duan Y, Ma G, Li X, Shen Z, Zhang S, Luo A, Zhong Z. Social media use, eHealth literacy, knowledge, attitudes, and practices toward COVID-19 vaccination among Chinese college students in the phase of regular epidemic prevention and control: a cross-sectional survey. Front Public Health. 2021;9:754904. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 11. Wang L, Li T, Gu R, Feng C. Large-scale meta-analyses and network analyses of neural substrates underlying human escalated aggression. NeuroImage. 2024;299:120824. [ DOI ] [ PubMed ] [ Google Scholar ] 12. Wang J, Zhang H, Tian D, An Y, Wang Y, Zheng Y. Examining the mediating effect of real and cyber self-control on cyberbullying and health risk behaviors among secondary vocational students in China: a cross-sectional study. BMC Public Health. 2024;24(1):2926. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 13. Sun M, Ma Z, Xu B, Chen C, Chen QW, Wang D. Prevalence of cyberbullying involvement and its association with clinical correlates among Chinese college students. J Affect Disord. 2024;367:374–81. [ DOI ] [ PubMed ] [ Google Scholar ] 14. Alghamdi W, Almadani S, Banjer H, Alsulami D, Alghamdi Y. Relationship between cyberbullying, anxiety, and depression among university students in Jeddah, Saudi Arabia: A cross-sectional study. Int J Psychiatry Med. 2025;60(2):170–86. [ DOI ] [ PubMed ] [ Google Scholar ] 15. Kim M, Ellithorpe M, Burt SA. Anonymity and its role in digital aggression: A systematic review. Aggress Violent Behav. 2023;72(000):11. [ Google Scholar ] 16. Thompson TJ, Munjal S. Discord and distress: cyberbullying and adolescent mental health. Prim Care Companion CNS Disord. 2024;26(6):24cr03797. [ DOI ] [ PubMed ] [ Google Scholar ] 17. Allen JP, Costello MA, Hunt GL, Uchino BN, Sugden K. Predictions from early adolescent interpersonal aggression to accelerated aging in adulthood: Relational and biological mechanisms of linkage. Health Psychol. 2026. 10.1037/hea0001576. [ DOI ] [ PMC free article ] [ PubMed ] 18. Abudawoud DH, Albukhari AF. Prevalence of internet gaming disorder and its association with aggression: an online survey among Saudi adolescents. BMC Psychol. 2026;14(1):200. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 19. Russell NM, Frier MD, Bortolato M, Mangieri RA. Decoding the bidirectional links between alcohol misuse and aggression: toward a unified translational framework. Psychopharmacology. 2025. 10.1007/s00213-025-06938-0. [ DOI ] [ PMC free article ] [ PubMed ] 20. Nagata JM, Trompeter N, Singh G, Ganson KT, Testa A, Jackson DB, Assari S, Murray SB, Bibbins-Domingo K, Baker FC. Social epidemiology of early adolescent cyberbullying in the United States. Acad Pediatr. 2022;22(8):1287–93. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 21. Jin X, Zhang K, Twayigira M, Gao X, Xu H, Huang C, Luo X, Shen Y. Cyberbullying among college students in a Chinese population: Prevalence and associated clinical correlates. Front Public Health. 2023;11:1100069. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 22. Anderson CA, Bushman BJ. Human aggression. Annu Rev Psychol. 2002;53(1):27–51. [ DOI ] [ PubMed ] [ Google Scholar ] 23. Altman DG, Royston P. The cost of dichotomising continuous variables. BMJ. 2006;332(7549):1080. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 24. Park H, Yoon EL, Kim M, Lee J, Kim HL, Cho S, Nah EH, Jun DW. Diagnostic performance of the fibrosis-4 index and the NAFLD fibrosis score for screening at-risk individuals in a health check-up setting. Hepatol Commun. 2023;7(10):e0249. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 25. Linden A, Yarnold PR. Using classification tree analysis to generate propensity score weights. J Eval Clin Pract. 2017;23(4):703–12. [ DOI ] [ PubMed ] [ Google Scholar ] 26. Hill RM, Oosterhoff B, Kaplow JB. Prospective identification of adolescent suicide ideation using classification tree analysis: Models for community-based screening. J Consult Clin Psychol. 2017;85(7):702–11. [ DOI ] [ PubMed ] [ Google Scholar ] 27. Zhang Y, Shi P, Gao M, Chang H. Factors Influencing Aggressive Adolescent Behavior: An Analysis Using the Decision Tree Method. J Genet Psychol. 2022;183(6):537–48. [ DOI ] [ PubMed ] [ Google Scholar ] 28. Hevey D. Network analysis: a brief overview and tutorial. Health Psychol Behav Med. 2018;6(1):301–28. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 29. Jones PJ, Ma R, McNally RJ. Bridge Centrality: A Network Approach to Understanding Comorbidity. Multivar Behav Res. 2021;56(2):353–67. [ DOI ] [ PubMed ] [ Google Scholar ] 30. Tian X, Zhao Y, Fan Q, Zhu W. Unveiling the Dual Nature of Aggression: A Network Analysis Approach. Int J Psychol. 2025;60(6):e70131. [ DOI ] [ PubMed ] [ Google Scholar ] 31. Chen L, Luo S, Cao J, Chen L, Ma T. A Network Analysis of College Students’ Coping Styles and Aggressive Behavior. Psychol Rep. 2025:332941251390457. 10.1177/00332941251390457. [ DOI ] [ PubMed ] 32. Zhu L, Huang M, Fang Z, Tong J, Pan Z, Hua L, Dong P, Wen L, Chang W, Yao Y, et al. Exploring the Relationship Between Aggressive Behavior, Family Parenting Styles, and Self-Esteem Among Only-Child College Students in China: A Cross-Sectional Study. Psychol Res Behav Manag. 2025;18:435–48. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 33. Feng Z, Wen-Bin G. Reliability and validity of the Adolescent Online Aggressive Behavior Scale. Chin Ment Health J. 2012;26(06):439–44. [ Google Scholar ] 34. Zheng Q, Ye B, Ye L, Guo S, Liao Y, Liu M. Effects of moral disengagement on online aggressive behavior in college students: The mediating role of moral identity and the moderating role of gender. Chin J Clin Psychol. 2016;24(04):714–6. [ Google Scholar ] 35. Jin T, Lu G, Zhang L, Wei L, Ma X, Suo H. Psychological maltreatment and cyber victimization in adolescents: mediating of social anxiety. Chin J Clin Psychol. 2017;25(01):167–70. [ Google Scholar ] 36. Jin T. The relationship between violence exposure and online aggressive behavior in college students: the role of ruminative responses and internet moral [Master’s thesis]. Harbin: Harbin Normal University; 2018. 37. Jin T, Wuyuntena, Zhang L, Lei Z, Jia Y. The transition of online aggressive behavior among college students: a latent transition analysis. Psychol Sci. 2024;47(03):614–21. [ Google Scholar ] 38. van Borkulo CD, Borsboom D, Epskamp S, Blanken TF, Boschloo L, Schoevers RA, Waldorp LJ. A new method for constructing networks from binary data. Sci Rep. 2014;4:5918. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 39. Brusco MJ, Steinley D, Hoffman M, Davis-Stober C, Wasserman S. On Ising models and algorithms for the construction of symptom networks in psychopathological research. Psychol Methods. 2019;24(6):735–53. [ DOI ] [ PubMed ] [ Google Scholar ] 40. Linden A, Yarnold PR. Identifying causal mechanisms in health care interventions using classification tree analysis. J Eval Clin Pract. 2018;24(2):353–61. [ DOI ] [ PubMed ] [ Google Scholar ] 41. Epskamp S, Borsboom D, Fried EI. Estimating psychological networks and their accuracy: A tutorial paper. Behav Res Methods. 2018;50(1):195–212. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 42. Haslbeck JMB, Waldorp LJ. How well do network models predict observations? On the importance of predictability in network models. Behav Res Methods. 2018;50(2):853–61. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 43. Chen S, Wei M, Wang X, Liao J, Li J, Liu Y. Competitive video game exposure increases aggression through impulsivity in Chinese adolescents: evidence from a multi-method study. J Youth Adolesc. 2024;53(8):1861–74. [ DOI ] [ PubMed ] [ Google Scholar ] 44. Wei MC, Chen S, Wang X, Liu YL. The impact of competitive gaming context on aggressive behavior in high school students: the positive role of cooperative gaming context. J Adolesc. 2025;97(6):1656–66. [ DOI ] [ PubMed ] [ Google Scholar ] 45. Teague S, Somoray K, Shatte A, Miller D, Moss K, Crawford A, Wildman H, Kayal D, Hutchinson D. Digital Media Use and Child Health and Development: A Systematic Review and Meta-Analysis. JAMA Pediatr. 2026. 10.1001/jamapediatrics.2026.0085. [ DOI ] [ PMC free article ] [ PubMed ] 46. Deng X, Hu YB, Liu CY, Li Q, Yang N, Zhang QY, Liu L, Qiu JN, Xu HB, Xue L, et al. Psychological distress and aggression among adolescents with internet gaming disorder symptoms. Psychiatry Res. 2024;331:115624. [ DOI ] [ PubMed ] [ Google Scholar ] 47. Zheng X, Chen H, Wang Z, Xie F, Bao Z. Online violent video games and online aggressive behavior among Chinese college students: The role of anger rumination and self-control. Aggress Behav. 2021;47(5):514–20. [ DOI ] [ PubMed ] [ Google Scholar ] 48. Dickmeis A, Roe K. Genres matter: Video games as predictors of physical aggression among adolescents. Communications. 2019;44(1):105–29. [ Google Scholar ] 49. Kim M. The faceless: anonymity and personality in digital aggression [Master’s thesis]. East Lansing (MI): Michigan State University; 2023. 50. Chen S, Hong H, Zhou Y, Huang X, Gao X. Neurobiological correlates of reactive aggression in young adults with internet gaming disorder. Brain Res Bull. 2025;220:111133. [ DOI ] [ PubMed ] [ Google Scholar ] 51. Li S, Wu Z, Zhang Y, Xu M, Wang X, Ma X. Internet gaming disorder and aggression: A meta-analysis of teenagers and young adults. Front Public Health. 2023;11:1111889. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 52. Ritter K, Lookatch SJ, Schmidt MR, Moore TM. The impact of history of aggression and alcohol use on aggressive responding in the laboratory. Subst Abuse. 2019;13:1178221819884328. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 53. Crane CA, Umehira N, Berbary C, Easton CJ. Problematic alcohol use as a risk factor for cyber aggression within romantic relationships. Am J Addict. 2018. 10.1111/ajad.12736. [ DOI ] [ PubMed ] 54. Sontate KV, Rahim Kamaluddin M, Naina Mohamed I, Mohamed RMP, Shaikh MF, Kamal H, Kumar J. Alcohol, Aggression, and Violence: From Public Health to Neuroscience. Front Psychol. 2021;12:699726. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 55. Tobore TO. On the Neurobiological Role of Oxidative Stress in Alcohol-Induced Impulsive, Aggressive and Suicidal Behavior. Subst Use Misuse. 2019;54(14):2290–303. [ DOI ] [ PubMed ] [ Google Scholar ] 56. Xiao B, Parent N, Bond T, Sam J, Shapka J. Developmental Trajectories of Cyber-Aggression among Early Adolescents in Canada: The Impact of Aggression, Gender, and Time Spent Online. Int J Environ Res Public Health. 2024;21(4):429. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 57. Yamada M, Sekine M, Tatsuse T, Asaka Y. Prevalence and Associated Factors of Pathological Internet Use and Online Risky Behaviors Among Japanese Elementary School Children. J Epidemiol. 2021;31(10):537–44. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 58. Szukało A, Klimek J. Rola interpersonalnego kontekstu rodzicielstwa i spostrzeganego wsparcia społecznego w regulacji agresji interpersonalnej. Psychologia Wychowawcza. 2024;72(30):56–70. [ Google Scholar ] 59. Wang X, Zhu N, Wei M, Chen S, Liu W, Liu Y. Longitudinal Associations and Gender Differences of Parent-Child Relationships on Aggression in Chinese Adolescents: The Mediating Role of Psychological Suzhi. Aggress Behav. 2025;51(3):e70031. [ DOI ] [ PubMed ] [ Google Scholar ] 60. Liu X, Zhang Y, Zeng J, Jiang Z, Liu Y. Development of Father- and Mother-Child Intimacy and Their Association with Internalizing and Externalizing Problems among Early and Middle Chinese Adolescents. J Youth Adolesc. 2025;54(5):1326–39. [ DOI ] [ PubMed ] [ Google Scholar ] 61. Chen M, Tian N. Associations between childhood victimization, parent-child closeness, and adulthood well-being outcomes: A network analysis. Child Abuse Negl. 2026;172:107866. [ DOI ] [ PubMed ] [ Google Scholar ] 62. Savage J, Ferguson CJ, Flores L. The effect of academic achievement on aggression and violent behavior: A meta-analysis. Aggress Violent Behav. 2017;37:91–101. [ Google Scholar ] 63. Kruglanski AW, Ellenberg M, Szumowska E, Molinario E, Speckhard A, Leander NP, Pierro A, Di Cicco G, Bushman BJ. Frustration–aggression hypothesis reconsidered: The role of significance quest. Aggress Behav. 2023;49(5):445–68. [ DOI ] [ PubMed ] [ Google Scholar ] 64. Granvik Saminathen M, Plenty S, Modin B. The Role of Academic Achievement in the Relationship between School Ethos and Adolescent Distress and Aggression: A Study of Ninth Grade Students in the Segregated School Landscape of Stockholm. J Youth Adolesc. 2021;50(6):1205–18. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 65. Buckeridge DL. Precision, Equity, and Public Health and Epidemiology Informatics - A Scoping Review. Yearb Med Inf. 2020;29(1):226–30. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 66. Roberts MC, Holt KE, Del Fiol G, Baccarelli AA, Allen CG. Precision public health in the era of genomics and big data. Nat Med. 2024;30(7):1865–73. [ DOI ] [ PMC free article ] [ PubMed ] [ Google Scholar ] 67. Seigfried-Spellar KC, Treadway KN. Differentiating hackers, identity thieves, cyberbullies, and virus writers by college major and individual differences. Deviant Behav. 2014;35(10):782–803. [ Google Scholar ] 68. Kunhao Z, Ma’rof AA, Abdullah H. The role of moral disengagement, anonymity perception, online disinhibition, and empathy in predicting cyberbullying perpetration among Chinese young adults. Int J Acad Res Bus Soc Sci. 2024;14(12):1114–27. [ Google Scholar ] 69. Nurdin MNH, Firdaus F, Fajrianti AA, Fadilah N, Piara M. Effect of Online Game’s Intensity on Verbal Aggression Behavior in students of SMP Negeri 10 Makassar. ARRUS J Social Sci Humanit. 2024;4(3):319–28. [ Google Scholar ] 70. Koanda N, Rahayu P, Zainuddin NI. Overview of Aggressiveness in Early Adolescents Who Play Online Games. Int J Educ Social Stud Manage (IJESSM). 2024;4(2):689–99. [ Google Scholar ] 71. Bringmann LF, Elmer T, Epskamp S, Krause RW, Schoch D, Wichers M, Wigman JTW, Snippe E. What do centrality measures measure in psychological networks? J Abnorm Psychol. 2019;128(8):892–903. [ DOI ] [ PubMed ] [ Google Scholar ] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Supplementary Material 1. (23.7KB, docx) Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Articles from BMC Public Health are provided here courtesy of BMC ACTIONS View on publisher site PDF (2.4 MB) Cite Collections Permalink RESOURCES Similar articles Cited by other articles Links to NCBI Databases Follow NCBI NCBI on X (formerly known as Twitter) NCBI on Facebook NCBI on LinkedIn NCBI on GitHub NCBI RSS feed Connect with NLM NLM on X (formerly known as Twitter) NLM on Facebook NLM on YouTube National Library of Medicine 8600 Rockville Pike Bethesda, MD 20894 Web Policies FOIA HHS Vulnerability Disclosure Help Accessibility Careers NLM NIH HHS USA.gov Back to Top

Related documents

• Nonlinear associations between punishment sensitivity and problematic Internet use among adolescents: Results from cross-sectional and longitudinal analyses.#170560
• Medium-specific escape motivation in internet gaming disorder and problematic social media use among Chinese university students: the role of self-control.#678275
• Examining the Role of Self-Construal on Cybersecurity Attitudes and Perceptions of Vulnerability#1120045
• Cybersecurity Vulnerability Behavior Scale in College During the Covid-19 Pandemic#672009
• A micro-sociological approach to prolonged grief disorder: identification and measurement of simpatico, a novel interpersonal risk factor#657788