Mapping Classrooms and Peer Victimization Using a Latent Profile Approach: Network, Normative and Demographic Factors

Mapping Classrooms and Peer Victimization Using a Latent Profile Approach: Network, Normative and Demographic Factors | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Mapping Classrooms and Peer Victimization Using a Latent Profile Approach: Network, Normative and Demographic Factors Beatriz Franco-Ugidos, Alar Urruticoechea, Elena Vernazza-Mañan, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7678286/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background: Bullying is a complex group phenomenon, influenced by classroom social conditions that cannot be understood solely through individual characteristics. While previous research has examined factors like cohesion, hierarchy, norms, and structural characteristics separately as gender ratio and class size little is known about how these elements combine to form distinct social configurations associated with different bullying levels. This study aimed to identify latent classroom profiles based on multiple social and structural indicators and examine their relationship with peer victimization patterns. Methods: A latent class analysis was conducted with 19,708 students from 746 classrooms across primary and secondary education in Spain, participating in the Sociescuela program. Data were collected through computer-based sociometric peer nominations assessing victimization, cohesion, hierarchy, aggressive norms, class size, and gender composition. Classroom-level indicators were standardized and analyzed using Gaussian Mixture Models. Model selection was based on multiple fit indices including AIC, BIC, aBIC, entropy, and likelihood ratio tests. Results: A three-class solution provided the optimal balance between statistical fit and interpretability. A female-dominated low-risk profile (Class 1, n = 80 classrooms, 11.6%) showed the lowest victimization levels (M = -0.40), below-average aggressive norms (M = -0.12), low cohesion (M = -0.59) and gender imbalance favoring girls (M = 1.26). A high-risk vulnerable profile (Class 2, n = 130 classrooms, 18.9%) exhibited the highest victimization levels (M = 0.56), the highest aggressive norms (M = 0.38), the smallest class sizes (M = -0.98), and high cohesion (M = 0.17), representing the most problematic classroom environment. A normative balanced profile (Class 3, n = 478 classrooms, 69.5%) demonstrated slightly below-average victimization (M = -0.09), moderate positive cohesion (M = 0.19), larger class sizes (M = 0.32), and below-average aggressive norms (M = -0.16), representing the most stable and typical classroom climate. Conclusions: The study identified three distinct classroom social configurations with implications for bullying prevention. Findings emphasize that victimization risk depends on combinations of social and structural factors. Class 2 represents the highest-risk environment requiring intensive intervention, while Classes 1 and 3 show different protective mechanisms. Results support intervention strategies tailored to specific classroom profiles rather than applying universal approaches. Bullying Peer victimization Latent class analysis Cohesion Hierarchy Classroom norms Gender ratio Class size Figures Figure 1 Figure 2 Figure 3 Figure 4 1. Introduction Bullying represents one of the most pervasive threats to student well-being in schools worldwide. International studies have consistently shown that bullying leads to severe physical, social, and emotional consequences for victims, perpetrators, and bystanders [ 1 ]. In 2019, UNESCO reported that nearly 32% of students had been bullied by their peers at least once during the month leading up to the measurement [ 2 ]. It is defined as an abuse of power among peers, characterized by aggressive behavior—physical, verbal, or psychological— deliberately exerted by the aggressor toward a victim who is unable to defend themselves effectively [ 3 ]. Bullying is a group phenomenon strongly influenced by the social conditions of the classroom. Its prevalence cannot be understood solely through the individual characteristics of those involved but depends on contextual factors that shape the social climate of the group [ 4 , 5 ]. These factors include shared attitudes toward aggression, interpersonal relationships, and other group dynamics that can either foster or inhibit bullying behaviors among peers. 1.1 Peer network structure Structural and normative classroom conditions strongly influence the dynamics of bullying. In this sense, one of the central components of classroom climate is group cohesion, which refers to the degree of unity and sense of belonging among students. More cohesive classrooms tend to exhibit less victimization, better student socioemotional adjustment, and a culture of greater collaboration [ 6 ]. However, cohesion is not always protective. When it is based on norms that tolerate or promote aggression, it may legitimize bullying behaviors and violence against those who deviate from group [ 7 ]. The social hierarchy among students, understood as differences in status in terms of leadership, popularity, or centrality, can generate power asymmetries [ 8 ]. In particular, high-status students who exhibit aggressive behaviors tend to maintain their position within the group, especially when their peers tolerate or even reward them [ 9 , 10 ]. Although classroom cohesion is often viewed as protective, its interaction with strong hierarchical structures can paradoxically exacerbate peer victimization. In highly cohesive classrooms where social bonds are tightly knit, students may become more embedded in peer networks that reinforce existing power asymmetries. When these networks are also hierarchical, they often elevate aggressive students while those who occupy marginal positions or have peripheral relationships are often more vulnerable to victimization. Ahn et al. [ 11 ] found that aggressive students achieved higher perceived popularity in classrooms characterized by both high density and strong embeddedness, whereas victimized students were most disliked in hierarchical environments. Martín Babarro et al. [ 12 ] further demonstrated that in Spanish secondary classrooms, the coexistence of cohesion and hierarchy predicted higher levels of victimization, with peer rejection being more strongly linked to victim status under these conditions. In such environments, deviant or vulnerable students are more likely to be excluded and less likely to be defended, as norms of group alignment suppress dissent. 1.2 Group norms These dynamics are closely tied to the normative system of the classroom, where group norms act as behavioral guides. In contexts where aggressive norms prevail, such behavior can reinforce the bully's social position and status [ 13 ]. The classroom social climate, understood as the set of explicit and implicit norms shared by students, plays a key role in the evolution of bullying [ 14 , 4 ]. Each classroom develops a system of rules indicating which behaviors are accepted and sanctioned [ 13 ]. When group norms value aggression or humiliation of others, bullying is likely to be promoted. Conversely, in environments where prosocial behaviors are encouraged, a lower prevalence of bullying is observed [ 5 ]. This normative influence also shapes bystander behavior: in classrooms where observers remain passive, bullying is reinforced, whereas when students support victims or actively reject aggression, perpetrators receive less social validation and reduce the recurrence of their attacks [ 16 ]. Adding to this, Aguilar-Pardo et al. [ 15 ] showed that network density amplifies the influence of dominant peer norms (whether prosocial or aggressive) and that when aggressive behaviors are endorsed by high-status individuals, victims experience sharper declines in peer acceptance. 1.3 Class size and gender ratio Contrary to common assumptions, Coelho and Romão [ 17 ] found that small classrooms sometimes increase victimization risk, particularly when social norms are aggressive and peer structures rigid. Similarly, research conducted in the Netherlands has reported a negative association between classroom size and peer-reported bullying, suggesting that larger classrooms may dilute the social power of coercive leaders and promote relational diversity [ 18 ]. Although small class sizes are often assumed to facilitate management and reduce aggression, research shows that they may sometimes increase the risk of victimization, particularly when social norms are aggressive and peer structures are rigid [ 17 ]. Gender composition of classrooms represents one of the least studied structural factors in bullying research, despite its relevance. While robust evidence on gender differences in victimization and aggression patterns [ 19 , 20 ], studies on how the proportion of boys and girls in classrooms influences bullying dynamics are extremely scarce. Busching and Krahé [ 21 ] found that girls' norms regarding aggression significantly influence the entire classroom climate, highlighting the role of gendered moral norms. Additionally, research suggests that status-driven aggression among boys may be reinforced by social rewards, especially in classrooms where dominance behaviors are tolerated [ 4 ]. 1.4 The present study Taken together, cohesion, hierarchy, group norms, class size, and gender composition constitute a framework that may define each classroom climate. While each of these variables has been studied individually, little is known about how they combine to create latent patterns of social organization associated with varying levels of bullying risk. Latent class or profile analysis (LCA/LPA) have revealed multiple bullying profiles—such as victims, bully-victims, defenders, or uninvolved students—and their transitions over time [ 22 , 23 , 24 , 25 ]. Individual, peer, and school-level factors—such as teacher and peer support, frequency of victimization, and sociodemographic characteristics—are key predictors of membership in these profiles [ 26 , 27 ]. These findings support the need for tailored interventions, as students' experiences with bullying vary by profile [ 28 ]. Understanding classroom profiles not only advances theoretical insight but also has practical relevance, enabling the design of differentiated anti-bullying strategies suited to each classroom configuration. This study addresses this gap through a latent profile approach, aiming to identify distinct profiles of social configuration that may be more or less conducive to peer victimization. This study aims to understand how classroom social configurations are organized into distinct latent patterns based on multiple indicators: peer victimization, group cohesion, status hierarchy, aggressive norms, group size, and gender composition. By identifying these profiles, the study seeks to inform the design of more effective and context-sensitive bullying prevention strategies. Intervention programs could be tailored to the predominant social characteristics of each classroom, allowing for a more efficient allocation of resources according to the identified profile [ 29 , 4 ]. 2. Methods 2.1 Participants This study was a cross-sectional design with non-probability sampling. The sample included 19,708 students in 746 classrooms from 70 schools (63 public and 7 blended schools), all of whom were enrolled in the Sociescuela program in Spain [ 30 ] ---an initiative aimed at assessing and improving classroom social dynamics through peer-reported data. Of the total sample, 9,079 were girls (47.7%) and 9,968 were boys (52.3%). Regarding educational level, 5,729 (29.1%) were students from primary education, and 13,964 (70.9%) were students from secondary education. 2.2 Procedure Data collection occurred during regular school hours across multiple academic periods from October 2022 to May 2023. All data collection sessions followed a standardized classroom procedure conducted in computer rooms over -minute periods. Only students who provided assent to participate and whose parents had given informed consent were included in the study. During each session, two trained research assistants presented the activity and provided standardized instructions for completing the questionnaires. The assistants emphasized that all responses would be treated with strict confidentiality and used exclusively for research purposes to improve educational processes. Participants completed online-based questionnaires (detailed in [ 30 ]) as part of larger-scale surveys designed to assess different levels related to school violence using the Sociescuela program. All measures were based on peer nominations within classrooms, allowing students to nominate any classmate by selecting from a list displaying the classmates' names. This procedure enabled participants to nominate even absent classmates, ensuring complete data collection for all students. The methodology allowed for the calculation of indices of aggression, cohesion, hierarchy and victimization for all students, with no missing data reported across the assessment periods. 2.3 Measures Peer Victimization was measured using a peer-nomination technique without restricting the number of nominations allowed. This method produced four types of victimization scores: physical ("Which of your classmates are often pushed around or beaten by other students? "), verbal ("Which of your classmates are regularly made fun of or insulted?"), relational ("Which of your classmates are usually ignored or ostracized?") and cyberbullying (Which of your classmates is bothered through mobile phones or on social media?). For each item, the number of nominations received by a student was divided by the number of students who responded to that item. These standardized scores were then averaged across the three victimization types to yield a composite victimization index (Cronbach's α = .83). The resulting index ranged from 0 to 0.57 (M = 0.01, SD = 0.04) and was subsequently z-standardized. Aggressive Class Norm. It was measured through three peer-nomination questions: which classmates (1) treated others poorly, (2) bothered their peers, and (3) had poor relationships with teachers (Cronbach's α = .89), with a limit of three nominations per item. The number of nominations was divided by the number of respondents per question, and the three scores were averaged (ranged from 0.05 to 0.44; M = 0.21, SD = 0.51), followed by z-transformation. To assess class norm aggressiveness, the aggregated measures by classroom were computed. Cohesion Social . It was calculated to assess the average level of interpersonal connectivity within each group [ 31 ]. It was measured as the average number of positive friendship nominations received by students. Students could nominate up to nine classmates whom they considered friends. The total number of received nominations was divided by the maximum number of possible nominations per group. Higher scores reflect greater network cohesion. Group density scores ranged from 0.73 to 12.55 (M = 8.51, SD = 1.82) and were standardized using z-scores. Hierarchy. It represents the level of centrality within each group. It was measured as the average standard deviation of friendship nominations at the classroom level. Hierarchy scores ranged from 0.0 to 7.48, M = 3.83 SD = 0.95) Gender ratio . It was calculated by dividing the number of girls by the number of boys in the classroom. Higher gender ratio values reflect a lower presence of opposite-gender classmates (ranged from 0.14 to 12.0, M = 1.02, SD = 0.64). Class size. It refers to the total number of students enrolled in a given classroom. In the context of the Spanish education system, where this study was conducted, national regulations stipulate a legal maximum of 30 students per classroom in compulsory education. Consequently, class sizes in this sample reflect variability within this normative threshold. In the present study, class size was measured as a continuous variable (ranged from 10 to 35, M = 22.43, SD = 4.78). Educational level. It was coded as a binary variable (0 = primary education, 1 = secondary education) to examine differences in classroom dynamics across educational stages. 2.4 Data Analysis A Latent Class Analysis (LCA) was conducted to identify distinct classroom profiles based on their structural and normative characteristics. Since the focus of the analysis is on the group dimension of the classroom, the data were aggregated at the classroom level. To ensure the validity of the indicators at the group level, only classrooms with 10 or more students were included. Six indicators were considered for the analysis; classroom level of peer victimization, classroom cohesion, hierarchy, gender ratio, classroom size (number of students) and aggressive class norm. All indicators were transformed into standardized scores (Z-scores) to ensure comparability and prevent scale differences from influencing model estimation. All analyses were conducted in Python using the following packages: scikit-learn [ 32 ] for Gaussian Mixture Modeling, pandas [ 33 ] for data manipulation, NumPy [ 34 ] for numerical computations, SciPy [ 35 ] for statistical tests, and pingouin [ 36 ] for post-hoc analyses. Multiple models with between two and five latent classes were estimated, and the selection of the optimal model was based on a combination of statistical fit and theoretical relevance criteria. The indices considered included the Akaike Information Criterion (AIC), the Bayesian Information Criterion (BIC), the adjusted BIC (aBIC), the log-likelihood, the entropy, as well as specific likelihood tests such as the Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) and the Bootstrapped Likelihood Ratio Test (BLRT). The estimation process aimed to identify the model that best represented the latent structure of classrooms in terms of social configuration, considering both statistical fit and parsimony and the theoretical interpretability of the number of classes. Given substantial heterogeneity among latent classes—evidenced by significant differences in classroom characteristics (all F > 14.74, p 0.38)—a stratified analysis approach was adopted. For each latent class, multiple regression analyses examined victimization predictors (proportion of girls, class size, cohesion, hierarchy, and aggressive norms) using standardized variables. Heterogeneity was evaluated through ANOVA for mean differences, coefficient ranges across classes, and predictive capacity comparisons. 3. Results 3.1 Correlations and descriptive analysis Table 1 presents the correlation matrix and descriptive statistics for all study variables. The analysis reveals several significant associations that provide insights into the interrelationships among classroom. The strongest correlation in the matrix was observed between group size and aggressive classroom norms ( r = − .573, p < .01), indicating that smaller classrooms are significantly more likely to develop permissive attitudes toward aggression. This finding suggests an important structural constraint on normative development within classroom contexts. Cohesion and hierarchy demonstrated a substantial positive correlation ( r = .550, p < .01), indicating that classrooms with stronger social bonds tend to have more pronounced status hierarchies. This relationship suggests that social organization and group structure are closely intertwined in classroom contexts. Group size showed additional positive correlations with both cohesion ( r = .313, p < .01) and hierarchy ( r = .342, p < .01), suggesting that larger classrooms may facilitate both greater social connectedness and more differentiated status structures. Victimization showed its strongest positive association with aggressive classroom norms ( r = .493, p < .01), Victimization was also negatively correlated with group size ( r = − .378, p < .01), suggesting that smaller classrooms are associated with higher levels of victimization. Additionally, victimization showed a positive correlation with cohesion ( r = .174, p < .01) and a negative correlation with girls' ratio ( r = − .114, p < .01). Girls' ratio showed negative correlations with hierarchy ( r = − .109, p < .01), indicating that female-dominated classrooms tend to have less pronounced status differentials. However, girls' ratio showed no significant association with aggressive classroom norms ( r = .004, ns ), suggesting that gender composition influences social structure but not necessarily normative attitudes toward aggression. Table 1 Correlations and Descriptive Statistics of Study Variables Variable M SD 1 2 3 4 5 6 1. Victimization 0.01 0.04 — 2. Cohesion 8.51 1.82 .174** — 3. Hierarchy 3.83 0.95 -0.045 .550** — 4. Girls ratio 1.02 0.64 − .114** -0.059 − .109** — 5. Group Size 22.43 4.78 − .378** .313** .342** − .107** — 6. Aggressive class norm 0.21 0.51 .493** -0.051 − .191** 0.004 − .573** — * Descriptive statistics show original scale values before standardization. Correlations were computed using standardized variables. M = Mean; SD = Standard Deviation. ** p < .01. Table 2 Model Fit Indices for Latent Class Analysis Model AIC BIC aBIC Entropy LMR-LRT ( p value) BLRT ( p value) 2-class 11195.176 11598.812 11138.415 0.887 < 0.001 < 0.001 3-class 10405.375 11013.097 10320.913 0.942 < 0.001 < 0.001 4-class 10054.746 10866.554 9942.584 0.920 0.439 < 0.001 5-class 8820.069 9835.963 8680.207 0.981 0.034 0.012 Note. AIC = Akaike Information Criterion; BIC = Bayesian Information Criterion; aBIC = sample-size adjusted Bayesian Information Criterion; LMR-LRT = Lo-Mendell-Rubin Likelihood Ratio Test; BLRT = Bootstrap Likelihood Ratio Test. 3.2 Model selection Based on multiple fit indices, the 3-class model was selected as the optimal solution. The BIC showed its optimal point at 3 classes, with values increasing thereafter. The Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) was significant for 2-class and 3-class solutions ( p < .001) but became non-significant for the 4-class model ( p = .439), indicating that additional classes do not substantially improve fit. The 3-class model demonstrated excellent classification quality (entropy = 0.942) and maintained interpretability while balancing statistical fit and parsimony. Although 4- and 5-class models showed lower AIC values, convergence issues and potential overfitting made the 3-class solution the most defensible choice. The comparative analysis of the estimated models (Table 2 ) shows substantial improvements in the fit indicators with an increasing number of classes, albeit with diminishing returns. Moving from 2 to 3 classes, a considerable reduction in the information criteria is observed, with a ΔAIC of 789.801, a ΔBIC of 585.715, and a ΔaBIC of 817.502, as well as an improvement in entropy (Δ = 0.055). Furthermore, both the LMR-LRT and the BLRT are statistically significant, reinforcing the validity of the three-class model. The transition from 3 to 4 classes continues this trend with a ΔAIC of 350.629, a ΔBIC of 146.543, and a ΔaBIC of 378.329. However, a slight decrease in entropy is observed (Δ = -0.022), and the LMR-LRT is no longer significant, indicating that the improvement in fit does not justify the increase in model complexity (Fig. 1 ). These results suggest that the three-class model represents an adequate balance between statistical fit and parsimony, offering an interpretable and robust solution for describing social heterogeneity between classrooms. 3.3 Latent Profile Analysis Three distinct latent classes emerged from the analysis, demonstrating significant heterogeneity in classroom social configurations (Fig. 2 ). Class 1 ( n = 80 classrooms, 11.6%) showed the lowest victimization levels of all classes ( M = -0.40, SD = 0.33) and below-average aggressive norms ( M = -0.12, SD = 0.33). This class is characterized by notably low cohesion ( M = -0.59, SD = 1.04) and below-average hierarchy ( M = -0.39, SD = 0.92). The most defining feature of this class is an extreme gender imbalance strongly favoring girls (girls ratio M = 1.26, SD = 1.96), combined with below-average class size ( M = -0.70, SD = 0.78). These characteristics suggest an environment with minimal peer aggression but significant challenges in social organization and group cohesion. Class 2 ( n = 130 classrooms, 18.9%) emerges as the most vulnerable and high-risk profile. It is characterized by the highest victimization levels across all classes ( M = 0.56, SD = 0.75), moderate cohesion ( M = 0.17, SD = 0.95), and below-average hierarchy ( M = -0.23, SD = 1.01). These classrooms are the smallest among all classes ( M = -0.98, SD = 0.51) and show a slight gender imbalance toward boys (girls ratio M = -0.13, SD = 0.44). Class 2 exhibits the highest aggressive norms ( M = 0.38, SD = 0.36), creating a context of elevated social risk with substantial potential for peer conflict and bullying behaviors. Class 3 ( n = 478 classrooms, 69.5%) represents the most normative and balanced profile among all identified classes. This class is characterized by slightly below-average victimization levels ( M = -0.09, SD = 0.37), moderate positive cohesion ( M = 0.19, SD = 0.80), and slightly above-average hierarchy ( M = 0.12, SD = 0.92). The gender distribution shows a slight imbalance toward boys (girls ratio M = -0.15, SD = 0.38), while class size is considerably above the mean ( M = 0.32, SD = 0.62). Aggressive norms are below average ( M = -0.16, SD = 0.18), indicating an environment with lower prevalence of aggressive behaviors and representing the most stable classroom climate. 3.4 Analysis of Variance and Effect Sizes Analysis of variance revealed highly significant differences across all measured variables ( p < .001), with effect sizes ranging from moderate to very large (Table 3 ). Class size emerged as the most defining characteristic of the latent classes, F (2, 685) = 276.81, p < .001, η² = .447. Aggressive classroom norms showed comparable importance, F (2, 685) = 250.35, p < .001, η² = .422. These two dimensions collectively explained the majority of differences between latent classes. Victimization levels, representing the primary outcome variable, showed substantial between-class variation, F (2, 685) = 136.20, p < .001, η² = .284. Girls' ratio demonstrated strong differentiation between classes, F (2, 685) = 119.69, p < .001, η² = .259. Classroom cohesion and hierarchy showed smaller but significant contributions to class differentiation. Cohesion explained approximately 7.7% of between-class variance, F (2, 685) = 28.60, p < .001, η² = .077, while hierarchy contributed approximately 4.1% of variance, F (2, 685) = 14.74, p < .001, η² = .041. Table 3 Descriptive Statistics and Statistical Tests for Variables Across Latent Classes Variable Class 1 M (SD) Class 2 M (SD) Class 3 M (SD) F-Statistic Post Hoc Interpretation Victimization -0.40 (0.33) 0.56 (0.75) -0.09 (0.37) 136.20*** 2 > 3 > 1 Cohesion -0.59 (1.04) 0.17 (0.95) 0.19 (0.80) 28.60*** 2, 3 > 1 Hierarchy -0.39 (0.92) -0.23 (1.01) 0.12 (0.92) 14.74*** 3 > 1, 2 Girls ratio 1.26 (1.96) -0.13 (0.44) -0.15 (0.38) 119.69*** 1 > 2, 3 Class Size -0.70 (0.78) -0.98 (0.51) 0.32 (0.62) 276.81*** 3 > 1 > 2 Aggressive class norm -0.12 (0.33) 0.38 (0.36) -0.16 (0.18) 250.35*** 2 > 1, 3 *Note. M = Mean; SD = Standard deviation. Statistical significance was tested using one-way ANOVA. *** p < .001. F values correspond to p < .001 for all variables. 3.5 Stratified Regression Analysis for Victimization Regression models revealed markedly different predictive capacity across latent classes (Table 4 ). Class 2 showed the highest predictive capacity ( R ² = .354, adjusted R ² = .328), followed by Class 3 ( R ² = .200, adjusted R ² = .146) and Class 1 ( R ² = .130, adjusted R ² = .121). The R ² range across classes (0.224) indicates substantial heterogeneity in victimization mechanisms. Class 1 ( n = 80), characterized by high proportion of girls and small classes, victimization was primarily predicted by internal classroom factors. Aggressive norms were the strongest predictor (β = .105, R ² i = .118), followed by hierarchy (β = − .076, R ² i = .101). Crucially, class size showed no association with victimization (β = .001, R ² i = .034), suggesting that structural factors are less relevant than internal group dynamics in these classrooms. Class 2 , ( n = 130) showed the highest predictive capacity. Class size was the dominant predictor (β = − .451, R ² i = .158), indicating that smaller classes are strongly associated with greater victimization. Paradoxically, cohesion also predicted greater victimization (β = .351, R ² i = .046), suggesting that group cohesion may facilitate collective victimization dynamics in these classrooms. Class size showed completely different effects across classes: null in Class 1 (β = .001), very strong and negative in Class 2 (β = − .451), and weak in Class 3 (β = − .077). This extreme heterogeneity demonstrates that relationships between classroom structure and victimization are not uniform, but fundamentally depend on the classroom's latent profile. Class 3 (n = 478) with large classes and relatively low victimization, showed a limited predictive capacity. Cohesion was the strongest predictor (β = .099, R ² i = .030), though with weak effect. Table 4 Stratified Regression Analysis for Victimization by Latent Class Variable Class 1 Class 2 Class 3 β ( R ² i ) β ( R ² i ) β ( R ² i ) Girls ratio 0.010 (0.012) 0.072 (0.012) -0.015 (0.003) Class size 0.001 (0.034) -0.451** (0.158) -0.077 (0.043) Cohesion -0.035 (0.023) 0.351** (0.046) 0.099 (0.030) Hierarchy -0.076 (0.101) 0.018 (0.002) -0.022 (0.000) Aggressive class norm 0.105* (0.118) 0.011 (0.025) 0.062 (0.052) R ² .200 .354 .130 Adjusted R ² .146 .328 .121 Note. β = standardized regression coefficient; R ² i = individual R ² contribution. Class 1 = Internal Conflict profile ( n = 80); Class 2 = Structural Vulnerability profile ( n = 130); Class 3 = Resilient profile ( n = 478). |β| >0.1. ** |β| >0.2. 3.6 Educational Levels and Profiles Distribution Next, we did an analysis of the three profiles across educational stages (Figs. 3 and 4 ). Analysis of class distribution reveals important developmental patterns in classroom social configurations. Class 3 (normative profile) appears more prevalent in secondary education with higher victimization in primary education. Class 2 (high-risk profile) shows a similar presence in both educational stages with higher victimization in primary education. Finally, class 1 (female-dominated profile) is predominant in secondary education with similar levels of victimization in both educational levels, potentially reflecting developmental differences in how gender composition affects classroom dynamics. 4. Discussion The goal of this study was to identify distinct classroom social configurations based on group cohesion, status hierarchy, aggressive class norms, peer victimization, class size, and gender ratio. Using latent profile analysis, three distinct classroom profiles emerged. The emergence of three latent classes provides empirical support for the theoretical premise that classrooms function as complex social ecosystems with distinct organizational patterns [ 4 ]. Certain combinations of variables tend to appear together in consistent patterns, creating classrooms that are more or less conducive to bullying. The most predominant profile, representing nearly 70% of classrooms, suggests that most educational environments naturally develop protective characteristics against peer victimization. This finding aligns with Wölfer and Scheithauer [ 6 ], who noted that cohesion is generally associated with socioemotional adjustment and collaboration. A more vulnerable profile presents a pattern, characterized by high cohesion, aggressive norms, and paradoxically small class size. The stratified regression analysis revealed that this profile showed the highest predictive capacity (R² = .354), with class size emerging as the dominant predictor. This counterintuitive indicating that bullying can intensify when social networks are dense and opportunities to escape bullies' control are limited [ 37 ]. Particularly striking is the positive association between cohesion and victimization in these vulnerable contexts (β = .351), suggesting what Garandeau and Cillessen [ 38 ] described as "false cohesion", where group unity is achieved through shared targeting of victims rather than genuine social bonds. In these settings, bullies often occupy central positions within the classroom's social network, making it harder for other students to intervene [ 8 ]. This parallels Rambaran et al. [ 7 ], who warned that cohesion may become maladaptive when tied to aggressive group norms. In such contexts, bullies often occupy high-status positions [ 9 , 8 ], and peer reinforcement further stabilizes their dominance [ 10 ]. This results are consistent with the evidence provided that in classrooms with high network density and strong embeddedness [ 11 ] or hierarchy [ 12 ] victimized students are strongly disliked, highlighting how cohesion can magnify hierarchical inequalities. Regarding the class size the vulnerable latent profile also illustrates this mechanism, combining high cohesion with small class size and aggressive norms. This combination magnifies power asymmetries and reduces victims' ability to escape peer control, supporting Garandeau et al. [ 37 ], who found that victimization intensifies when networks are dense. It also mirrors findings from Coelho and Romão [ 17 ] and Fekkes et al. [ 18 ], who showed that smaller classes can heighten bullying risks when peer structures are rigid and coercive leaders dominate. The least common profile (female-dominated profile) characterized by gender imbalance favoring girls yet showing the lowest victimization levels, and with the lowest levels of social structure (cohesion and hierarchy). This configuration demonstrates that victimization was primarily predicted by aggressive norms (β = .105) while class size showed no association (β = .001), suggesting that internal normative processes rather than structural factors determine outcomes. The high female presence may offer a buffering effect against victimization, possibly due to different socialization processes and interactional styles [ 21 , 39 ]. The notably low cohesion indicates that protection from victimization does not necessarily translate to connectedness in the classroom. 4.1 Implications for Educational Practice These findings have significant implications for intervention. The variation in predictive patterns across profiles (R² range = 0.224) demonstrates that effective interventions must be tailored to specific classroom social configurations rather than applying universal approaches. For classrooms characterized by structural vulnerability, interventions should focus on restructuring social hierarchies and monitoring group dynamics. Special attention should be paid to the paradoxical role of cohesion in these contexts, ensuring that group unity is built around prosocial rather than exclusionary practices. Classrooms that demonstrate protective characteristics can benefit from approaches that reinforce and consolidate their existing strengths. Gender-imbalanced classrooms with low victimization require interventions that strengthen social cohesion while maintaining their naturally protective characteristics. Activities that promote inclusive group formation while leveraging the prosocial tendencies associated with higher female representation may be particularly effective. 4.2 Limitations and Future Directions Several limitations should be acknowledged. First, the cross-sectional design limits causal inferences about relationships between classroom characteristics and victimization outcomes. This prevents observation of how classroom dynamics evolve. Longitudinal studies are needed to establish whether classroom profiles represent stable configurations or dynamic states that change over time. Second, although the focus was on group characteristics, many bullying incidents relate to individual or family factors not considered in this study. Additionally, no information was collected on teacher behaviors or school-wide policies, both of which could influence classroom dynamics. Finally, although the sample is large and represents widely used programs in Spain, these profiles should be validated in other countries and educational contexts to assess their applicability across different environments 5. Conclusion This study demonstrates that classrooms are not uniform contexts for peer victimization but social systems with distinct organizational patterns. The identification of three distinct configurations provides a framework for understanding classroom-level heterogeneity in peer victimization processes and reinforces the consideration of bullying as a group-driven phenomenon embedded in peer dynamics and social norms. Such finding avoid the assumption that all classrooms require identical interventions. Abbreviations AIC - Akaike Information Criterion aBIC - Sample-size adjusted Bayesian Information Criterion ANOVA - Analysis of Variance BIC - Bayesian Information Criterion BLRT - Bootstrap Likelihood Ratio Test GMM - Gaussian Mixture Model LCA - Latent Class Analysis LMR-LRT - Lo-Mendell-Rubin Likelihood Ratio Test M - Mean SD - Standard Deviation Declarations Ethics Approval and Consent to Participate This study was approved by the Research Ethics Committee of Hospital General Universitario Gregorio Marañón (Code: Sociescuela-GM, Protocol version 1.0, dated February 21, 2025). The committee issued a favorable opinion on March 3, 2025 (Act 03/2025), certifying that the study follows legally established requirements and is pertinent for implementation. The study was approved for conduct without consent from source subjects for data utilization, as the research involves secondary analysis of anonymized educational data collected through the Sociescuela program. The committee confirmed that the protocol meets necessary requirements of suitability in relation to study objectives and that foreseeable risks and inconveniences to subjects are justified. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments. Consent for Publication Not applicable. This study involved secondary analysis of anonymized data collected through the Sociescuela program. No individual participant data that could lead to identification is presented in this manuscript. Availability of Data and Materials The datasets analyzed during the current study are available from the corresponding author on reasonable request, subject to appropriate ethical approval and data sharing agreements. Raw data cannot be shared publicly due to privacy restrictions and ethical considerations related to student information, but aggregated and anonymized datasets supporting the conclusions of this article may be made available to qualified researchers upon request. Competing Interests The authors declare that they have no competing interests. Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. The Sociescuela program data collection was supported by participating educational institutions. Authors' Contributions B.F.U. contributed to data collection, database processing, and manuscript writing. A.U. participated in database processing and manuscript writing. E.V.M. provided critical review and contributed to manuscript writing. V.S.L. designed the study, provided critical review and contributed to manuscript writing. J.M.B. (corresponding author) conducted data analysis, developed the methodology, supervised the study, and contributed to manuscript writing. All authors read and approved the final manuscript. Acknowledgements We acknowledge the participation of students, teachers, and educational institutions involved in the Sociescuela program. We thank the Research Ethics Committee of Hospital General Universitario Gregorio Marañón for their ethical review and approval of this study. Special appreciation goes to all the schools and educational professionals who facilitated data collection and supported this research initiative. We also thank the students who participated in the sociometric assessments that made this research possible Declaration of generative AI and AI-assisted technologies in the writing process. During the preparation of this work, the author(s) used DeepL Write and Grammarly to improve their English writing. After using this tool/service, the author(s) reviewed and edited the content as needed and took(s) full responsibility for the content of the published article. References Halliday S, Gregory T, Taylor A, Diggins E, Turnbull D. Adolescent mental health during the COVID-19 pandemic: A longitudinal study. J Affect Disord. 2021;290:336-45. UNESCO. Behind the numbers: Ending school violence and bullying. Paris: United Nations Educational, Scientific and Cultural Organization; 2019. Olweus D. Bullying at school: What we know and what we can do. Oxford: Blackwell Publishing; 1993. Salmivalli C. Bullying and the peer group: A review. Aggression Violent Behav. 2010;15(2):112-20. Laninga-Wijnen L, Steglich C, Harakeh Z, Vollebergh W, Veenstra R, Dijkstra JK. The role of prosocial and aggressive popularity norm combinations in prosocial and aggressive friendship processes. J Youth Adolesc. 2020;49(3):645-63. Wölfer R, Scheithauer H. Social influence and bullying behavior: Intervention-based network dynamics of the fairplayer.manual bullying prevention program. Aggressive Behav. 2014;40(4):309-19. Rambaran AJ, Hopmeyer A, Schwartz D, Steglich C, Badaly D, Veenstra R. Academic functioning and peer influences: A short-term longitudinal study of network-behavior dynamics in middle adolescence. Child Dev. 2020;88(2):523-43. Pouwels JL, Souren PM, Lansu TA, Cillessen AH. Stability of peer victimization: A meta-analysis of longitudinal research. Dev Rev. 2016;40:1-24. Dijkstra JK, Lindenberg S, Veenstra R. Beyond the class norm: Bullying behavior of popular adolescents and its relation to peer acceptance and rejection. J Abnorm Child Psychol. 2008;36(8):1289-99. Laninga-Wijnen L, Harakeh Z, Garandeau CF, Dijkstra JK, Veenstra R, Vollebergh WAM. Classroom popularity hierarchy predicts prosocial and aggressive popularity norms across the school year. Child Dev. 2019;90(5):e637-e655. Ahn HJ, Garandeau CF, Rodkin PC. Effects of classroom embeddedness and density on the social status of aggressive and victimized children. J Early Adolesc. 2010;30(1):76-101. Martín-Babarro J, Díaz-Aguado MJ, Martínez-Arias R, Steglich C. Power structure in the peer group: The role of classroom cohesion and hierarchy in peer acceptance and rejection of victimized and aggressive students. J Early Adolesc. 2017;37(9):1197-220. Saarento S, Salmivalli C. The role of classroom peer ecology and bystanders' responses in bullying. Child Dev Perspect. 2015;9(4):201-5. Saarento S, Kärnä A, Hodges EVE, Salmivalli C. Student-, classroom-, and school-level risk factors for victimization. J Sch Psychol. 2013;51(4):421-34. Aguilar-Pardo D, Martínez-Fernández B, Colmenares F, Martín-Babarro J. Peer likeability and victimization in young adolescents: Moderating effects of descriptive and status group norms for aggression and prosocial behaviour and network density. Br J Educ Psychol. 2022;92(3):925-54. Pozzoli T, Gini G. Active defending and passive bystanding behavior in bullying: The role of personal characteristics and perceived peer pressure. J Abnorm Child Psychol. 2010;38(6):815-27. Coelho VA, Romão AM. The relation between social anxiety, social withdrawal and (cyber)bullying roles: A multilevel analysis. Comput Human Behav. 2018;86:218-26. Fekkes M, Pijpers FI, Fredriks AM, Vogels T, Verloove-Vanhorick SP. Do bullied children get ill, or do ill children get bullied? A prospective cohort study on the relationship between bullying and health-related symptoms. Pediatrics. 2006;117(5):1568-74. Kowalski RM, Limber SP. Psychological, physical, and academic correlates of cyberbullying and traditional bullying. J Adolesc Health. 2013;53(1):S13-S20. Cross D, Shaw T, Hadwen K, Cardoso P, Slee P, Roberts C, et al. Longitudinal impact of the Cyber Friendly Schools program on adolescents' cyberbullying behavior. Aggressive Behav. 2022;42(2):166-80. Busching R, Krahé B. The girls set the tone: Gendered classroom norms and the development of aggression in adolescence. Pers Soc Psychol Bull. 2015;41(5):659-76. Lawson MA, Alameda-Lawson T, Downer J, Anderson E. Analyzing sub-population profiles and risk factors for school bullying. Child Youth Serv Rev. 2013;35(6):973-83. Pan Y, Liu H, Lau P, Luo F. A latent transition analysis of bullying and victimization in Chinese primary school students. PLoS One. 2017;12(8):e0182802. Liu J, Guo S, Weissman R, Liu H. Investigating factors associated with bullying utilizing latent class analysis among adolescents. Sch Psychol Int. 2021;42(1):11-32. Jenkins LN, Snyder Kaminski S, Miller M. Bystander Intervention in Bullying: Differences Across Latent Profiles. Int J Bullying Prev. 2021;3(2):130-7. Waasdorp TE, Bradshaw CP. Examining variation in adolescent bystanders' responses to bullying. School Psychol Rev. 2018;47(1):18-33. Ferreira-Junior V, Valente JY, Sanchez ZM. Examining associations between race, gender, alcohol use, school performance, and patterns of bullying in the school context: A latent class analysis. J Interpers Violence. 2021;37(15-16):NP12857-NP12880. Pivec T, Horvat M, Košir K. Psychosocial characteristics of bullying participants: A person-oriented approach combining self- and peer-report measures. Child Youth Serv Rev. 2023;144:106729. Gaffney H, Ttofi MM, Farrington DP. What works in anti-bullying programs? Analysis of effective intervention components. J Sch Psychol. 2021;85:37-56. Martín-Babarro J. Assessment and detection of peer-bullying through analysis of the group context. Psicothema. 2014;26(3):357-63. Wasserman S, Faust K. Social network analysis: Methods and applications. Cambridge: Cambridge University Press; 1994. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine learning in Python. J Mach Learn Res. 2011;12:2825-30. McKinney W. Data structures for statistical computing in Python. In: van der Walt S, Millman J, editors. Proceedings of the 9th Python in Science Conference; 2010. p. 51-6. Harris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D, et al. Array programming with NumPy. Nature. 2020;585(7825):357-62. Virtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: Fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17(3):261-72. Vallat R. Pingouin: Statistics in Python. J Open Source Softw. 2018;3(31):1026. Garandeau CF, Lee IA, Salmivalli C. Inequality matters: classroom status hierarchy and adolescents' bullying. J Youth Adolesc. 2019;48(6):1212-28. Garandeau CF, Cillessen AHN. From indirect aggression to invisible aggression: A conceptual view on bullying and peer group manipulation. Aggression Violent Behav. 2006;11(6):612-25. Coelho VA, Sousa V. Differential effectiveness of a middle school social and emotional learning program: does setting matter? J Youth Adolesc. 2018;47(9):1978-1991. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7678286","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":541897714,"identity":"15e7ab44-ba3c-4b90-95f1-de51db2b443c","order_by":0,"name":"Beatriz Franco-Ugidos","email":"","orcid":"","institution":"Complutense University of Madrid","correspondingAuthor":false,"prefix":"","firstName":"Beatriz","middleName":"","lastName":"Franco-Ugidos","suffix":""},{"id":541897715,"identity":"22913411-ce4a-4022-bbd6-664b3223dc4e","order_by":1,"name":"Alar Urruticoechea","email":"","orcid":"","institution":"Catholic University of Uruguay","correspondingAuthor":false,"prefix":"","firstName":"Alar","middleName":"","lastName":"Urruticoechea","suffix":""},{"id":541897716,"identity":"212572f4-9e7f-4f6a-947d-163591699e89","order_by":2,"name":"Elena Vernazza-Mañan","email":"","orcid":"","institution":"University of the Republic","correspondingAuthor":false,"prefix":"","firstName":"Elena","middleName":"","lastName":"Vernazza-Mañan","suffix":""},{"id":541897717,"identity":"fc9b4cec-3e6d-4be3-be82-650d138a623d","order_by":3,"name":"Vanesa Sainz","email":"","orcid":"","institution":"Francisco de Vitoria University","correspondingAuthor":false,"prefix":"","firstName":"Vanesa","middleName":"","lastName":"Sainz","suffix":""},{"id":541897718,"identity":"898cf4b9-c903-4c6b-aa11-8bae0694d953","order_by":4,"name":"Javier Martín-Babarro","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA3UlEQVRIiWNgGAWjYBACxhkMbGAGPxAfIE2LZAOxWhgkoFoMiFTPwMA8u/nZg5977skZ325/eICxzS6agb39AX6HzTlmbtjzrNjY7M4ZA6CW5NwGnjMG+LXMSDCT4DmQkLjtRg4DUMuB3AaJHPwOY5yR/k3yD1DL5hnpDyBa5J8TcNiMHDNpkC0bJBIMoLYwEHDYnDPlxjIHEowlQH5JOJec28aTg1+L4ez2bQ/fHEiQ45/d/vjDhzK73H724/gdZtgAY0kAcQIQs+FVDwTycJYEIaWjYBSMglEwYgEAXXtMQEgtvHgAAAAASUVORK5CYII=","orcid":"","institution":"Complutense University of Madrid","correspondingAuthor":true,"prefix":"","firstName":"Javier","middleName":"","lastName":"Martín-Babarro","suffix":""}],"badges":[],"createdAt":"2025-09-22 14:23:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7678286/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7678286/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":95542355,"identity":"5be3af43-5dd1-41d2-b765-3608d550491c","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":266890,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/231e575627a51353c6e8c178.docx"},{"id":95654721,"identity":"a2907be8-55f2-40e9-a39f-9d0a6c847244","added_by":"auto","created_at":"2025-11-11 16:12:50","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":8272,"visible":true,"origin":"","legend":"","description":"","filename":"677dd7b4a8604fc6a711bd28c74660c3.json","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/e3e12f57fd091da81e88f44a.json"},{"id":95654037,"identity":"600b2e98-4727-481c-90fc-8deed264790a","added_by":"auto","created_at":"2025-11-11 16:09:04","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124475,"visible":true,"origin":"","legend":"","description":"","filename":"677dd7b4a8604fc6a711bd28c74660c31enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/33693731bf011b2cfcaf5fc0.xml"},{"id":95654593,"identity":"25b49f01-19da-4c7f-aa1c-891ef2195899","added_by":"auto","created_at":"2025-11-11 16:12:32","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":50668,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/773bfc3ae0beabeda1786fd0.png"},{"id":95542358,"identity":"6e191b53-0245-44a1-b9a7-4a1c72e17186","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":97936,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/b8fd27f2cc83867d293aa285.png"},{"id":95542350,"identity":"28bf2396-1d86-4218-95d6-9e071058ea4b","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30736,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/26fedce7af17e0ce2ea467a2.png"},{"id":95542359,"identity":"3ec5c3a9-d696-4c4b-a421-ce453920e069","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29417,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/144c4ddafaeab7641de13865.png"},{"id":95542353,"identity":"fb8f599e-271d-408e-885b-b9624cc6fb18","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":23899,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/34f3b3665cf254261803fe67.png"},{"id":95542364,"identity":"38bede03-6709-4873-8942-4da0b9b8e62b","added_by":"auto","created_at":"2025-11-10 11:49:22","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":46749,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/a66ea8efd33f7351d85e93b3.png"},{"id":95654633,"identity":"79cdf669-6c3b-43a0-adc6-ca8262bf4f50","added_by":"auto","created_at":"2025-11-11 16:12:37","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14346,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/de253c15e3389b0bc55fe68b.png"},{"id":95654912,"identity":"6b6b9ec2-0ec9-476f-a3bc-5fbdcdbcb8e8","added_by":"auto","created_at":"2025-11-11 16:13:46","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12569,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/ab2281b07895724d78d8fbe8.png"},{"id":95542363,"identity":"8a859a35-694f-46c0-b7d6-fe354ce53129","added_by":"auto","created_at":"2025-11-10 11:49:22","extension":"xml","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122509,"visible":true,"origin":"","legend":"","description":"","filename":"677dd7b4a8604fc6a711bd28c74660c31structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/6bb28ca77ba1cebe269863db.xml"},{"id":95542366,"identity":"f04ebf3b-4e75-4d7d-ad4e-432620ad2c47","added_by":"auto","created_at":"2025-11-10 11:49:22","extension":"html","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133841,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/e0cf65e56ec058d06a23cad4.html"},{"id":95542348,"identity":"bad995ef-2215-4ff5-b2f7-4fd9412703e5","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":50668,"visible":true,"origin":"","legend":"\u003cp\u003eModel Fit Index by Number of Latent Clases\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/102e5b9550d28545105564a0.png"},{"id":95654504,"identity":"299fd1bd-5b04-4c95-896d-74f2993bfabc","added_by":"auto","created_at":"2025-11-11 16:12:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":97936,"visible":true,"origin":"","legend":"\u003cp\u003eMean Factor Scores by Latent Class Profile\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/8e17affe905f045d032a4fe6.png"},{"id":95654102,"identity":"552ac5ad-3dec-4208-9696-7d6578362a71","added_by":"auto","created_at":"2025-11-11 16:09:43","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":30736,"visible":true,"origin":"","legend":"\u003cp\u003eEducational Levels and Profiles Distribution\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/8b132a1dce22bb8fdb25a012.png"},{"id":95542346,"identity":"fbae4a35-4119-49d2-9234-909875982a3a","added_by":"auto","created_at":"2025-11-10 11:49:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":29417,"visible":true,"origin":"","legend":"\u003cp\u003eVictimization and Profiles Distribution by Educational Levels.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/bb65a32780fa91b339bc76d1.png"},{"id":101397956,"identity":"582a4668-4529-4da4-ac2c-068cd4a114c7","added_by":"auto","created_at":"2026-01-29 09:38:21","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1180094,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7678286/v1/f31d1277-8fe9-4707-9d8d-4a1ddc428942.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Mapping Classrooms and Peer Victimization Using a Latent Profile Approach: Network, Normative and Demographic Factors","fulltext":[{"header":"1. Introduction","content":"\u003cp\u003eBullying represents one of the most pervasive threats to student well-being in schools worldwide. International studies have consistently shown that bullying leads to severe physical, social, and emotional consequences for victims, perpetrators, and bystanders [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. In 2019, UNESCO reported that nearly 32% of students had been bullied by their peers at least once during the month leading up to the measurement [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. It is defined as an abuse of power among peers, characterized by aggressive behavior\u0026mdash;physical, verbal, or psychological\u0026mdash; deliberately exerted by the aggressor toward a victim who is unable to defend themselves effectively [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eBullying is a group phenomenon strongly influenced by the social conditions of the classroom. Its prevalence cannot be understood solely through the individual characteristics of those involved but depends on contextual factors that shape the social climate of the group [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. These factors include shared attitudes toward aggression, interpersonal relationships, and other group dynamics that can either foster or inhibit bullying behaviors among peers.\u003c/p\u003e\u003cdiv id=\"Sec2\" class=\"Section2\"\u003e\u003ch2\u003e1.1 Peer network structure\u003c/h2\u003e\u003cp\u003eStructural and normative classroom conditions strongly influence the dynamics of bullying. In this sense, one of the central components of classroom climate is group cohesion, which refers to the degree of unity and sense of belonging among students. More cohesive classrooms tend to exhibit less victimization, better student socioemotional adjustment, and a culture of greater collaboration [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. However, cohesion is not always protective. When it is based on norms that tolerate or promote aggression, it may legitimize bullying behaviors and violence against those who deviate from group [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. The social hierarchy among students, understood as differences in status in terms of leadership, popularity, or centrality, can generate power asymmetries [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. In particular, high-status students who exhibit aggressive behaviors tend to maintain their position within the group, especially when their peers tolerate or even reward them [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. Although classroom cohesion is often viewed as protective, its interaction with strong hierarchical structures can paradoxically exacerbate peer victimization. In highly cohesive classrooms where social bonds are tightly knit, students may become more embedded in peer networks that reinforce existing power asymmetries. When these networks are also hierarchical, they often elevate aggressive students while those who occupy marginal positions or have peripheral relationships are often more vulnerable to victimization. Ahn et al. [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e] found that aggressive students achieved higher perceived popularity in classrooms characterized by both high density and strong embeddedness, whereas victimized students were most disliked in hierarchical environments. Mart\u0026iacute;n Babarro et al. [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] further demonstrated that in Spanish secondary classrooms, the coexistence of cohesion and hierarchy predicted higher levels of victimization, with peer rejection being more strongly linked to victim status under these conditions. In such environments, deviant or vulnerable students are more likely to be excluded and less likely to be defended, as norms of group alignment suppress dissent.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e1.2 Group norms\u003c/h2\u003e\u003cp\u003eThese dynamics are closely tied to the normative system of the classroom, where group norms act as behavioral guides. In contexts where aggressive norms prevail, such behavior can reinforce the bully's social position and status [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. The classroom social climate, understood as the set of explicit and implicit norms shared by students, plays a key role in the evolution of bullying [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Each classroom develops a system of rules indicating which behaviors are accepted and sanctioned [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. When group norms value aggression or humiliation of others, bullying is likely to be promoted. Conversely, in environments where prosocial behaviors are encouraged, a lower prevalence of bullying is observed [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. This normative influence also shapes bystander behavior: in classrooms where observers remain passive, bullying is reinforced, whereas when students support victims or actively reject aggression, perpetrators receive less social validation and reduce the recurrence of their attacks [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. Adding to this, Aguilar-Pardo et al. [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e] showed that network density amplifies the influence of dominant peer norms (whether prosocial or aggressive) and that when aggressive behaviors are endorsed by high-status individuals, victims experience sharper declines in peer acceptance.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e1.3 Class size and gender ratio\u003c/h2\u003e\u003cp\u003eContrary to common assumptions, Coelho and Rom\u0026atilde;o [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] found that small classrooms sometimes increase victimization risk, particularly when social norms are aggressive and peer structures rigid. Similarly, research conducted in the Netherlands has reported a negative association between classroom size and peer-reported bullying, suggesting that larger classrooms may dilute the social power of coercive leaders and promote relational diversity [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Although small class sizes are often assumed to facilitate management and reduce aggression, research shows that they may sometimes increase the risk of victimization, particularly when social norms are aggressive and peer structures are rigid [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eGender composition of classrooms represents one of the least studied structural factors in bullying research, despite its relevance. While robust evidence on gender differences in victimization and aggression patterns [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e], studies on how the proportion of boys and girls in classrooms influences bullying dynamics are extremely scarce. Busching and Krah\u0026eacute; [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] found that girls' norms regarding aggression significantly influence the entire classroom climate, highlighting the role of gendered moral norms. Additionally, research suggests that status-driven aggression among boys may be reinforced by social rewards, especially in classrooms where dominance behaviors are tolerated [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e1.4 The present study\u003c/h2\u003e\u003cp\u003eTaken together, cohesion, hierarchy, group norms, class size, and gender composition constitute a framework that may define each classroom climate. While each of these variables has been studied individually, little is known about how they combine to create latent patterns of social organization associated with varying levels of bullying risk. Latent class or profile analysis (LCA/LPA) have revealed multiple bullying profiles\u0026mdash;such as victims, bully-victims, defenders, or uninvolved students\u0026mdash;and their transitions over time [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Individual, peer, and school-level factors\u0026mdash;such as teacher and peer support, frequency of victimization, and sociodemographic characteristics\u0026mdash;are key predictors of membership in these profiles [\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. These findings support the need for tailored interventions, as students' experiences with bullying vary by profile [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. Understanding classroom profiles not only advances theoretical insight but also has practical relevance, enabling the design of differentiated anti-bullying strategies suited to each classroom configuration.\u003c/p\u003e\u003cp\u003eThis study addresses this gap through a latent profile approach, aiming to identify distinct profiles of social configuration that may be more or less conducive to peer victimization. This study aims to understand how classroom social configurations are organized into distinct latent patterns based on multiple indicators: peer victimization, group cohesion, status hierarchy, aggressive norms, group size, and gender composition. By identifying these profiles, the study seeks to inform the design of more effective and context-sensitive bullying prevention strategies. Intervention programs could be tailored to the predominant social characteristics of each classroom, allowing for a more efficient allocation of resources according to the identified profile [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e].\u003c/p\u003e\u003c/div\u003e"},{"header":"2. Methods","content":"\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Participants\u003c/h2\u003e\u003cp\u003eThis study was a cross-sectional design with non-probability sampling. The sample included 19,708 students in 746 classrooms from 70 schools (63 public and 7 blended schools), all of whom were enrolled in the Sociescuela program in Spain [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e] ---an initiative aimed at assessing and improving classroom social dynamics through peer-reported data. Of the total sample, 9,079 were girls (47.7%) and 9,968 were boys (52.3%). Regarding educational level, 5,729 (29.1%) were students from primary education, and 13,964 (70.9%) were students from secondary education.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Procedure\u003c/h2\u003e\u003cp\u003eData collection occurred during regular school hours across multiple academic periods from October 2022 to May 2023. All data collection sessions followed a standardized classroom procedure conducted in computer rooms over -minute periods. Only students who provided assent to participate and whose parents had given informed consent were included in the study. During each session, two trained research assistants presented the activity and provided standardized instructions for completing the questionnaires. The assistants emphasized that all responses would be treated with strict confidentiality and used exclusively for research purposes to improve educational processes. Participants completed online-based questionnaires (detailed in [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]) as part of larger-scale surveys designed to assess different levels related to school violence using the Sociescuela program. All measures were based on peer nominations within classrooms, allowing students to nominate any classmate by selecting from a list displaying the classmates' names. This procedure enabled participants to nominate even absent classmates, ensuring complete data collection for all students. The methodology allowed for the calculation of indices of aggression, cohesion, hierarchy and victimization for all students, with no missing data reported across the assessment periods.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Measures\u003c/h2\u003e\u003cp\u003e\u003cem\u003ePeer Victimization\u003c/em\u003e was measured using a peer-nomination technique without restricting the number of nominations allowed. This method produced four types of victimization scores: physical (\"Which of your classmates are often pushed around or beaten by other students? \"), verbal (\"Which of your classmates are regularly made fun of or insulted?\"), relational (\"Which of your classmates are usually ignored or ostracized?\") and cyberbullying (Which of your classmates is bothered through mobile phones or on social media?). For each item, the number of nominations received by a student was divided by the number of students who responded to that item. These standardized scores were then averaged across the three victimization types to yield a composite victimization index (Cronbach's α\u0026thinsp;=\u0026thinsp;.83). The resulting index ranged from 0 to 0.57 (M\u0026thinsp;=\u0026thinsp;0.01, SD\u0026thinsp;=\u0026thinsp;0.04) and was subsequently z-standardized.\u003c/p\u003e\u003cp\u003e\u003cem\u003eAggressive Class Norm.\u003c/em\u003e It was measured through three peer-nomination questions: which classmates (1) treated others poorly, (2) bothered their peers, and (3) had poor relationships with teachers (Cronbach's α\u0026thinsp;=\u0026thinsp;.89), with a limit of three nominations per item. The number of nominations was divided by the number of respondents per question, and the three scores were averaged (ranged from 0.05 to 0.44; M\u0026thinsp;=\u0026thinsp;0.21, SD\u0026thinsp;=\u0026thinsp;0.51), followed by z-transformation. To assess class norm aggressiveness, the aggregated measures by classroom were computed.\u003c/p\u003e\u003cp\u003e\u003cem\u003eCohesion Social\u003c/em\u003e. It was calculated to assess the average level of interpersonal connectivity within each group [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. It was measured as the average number of positive friendship nominations received by students. Students could nominate up to nine classmates whom they considered friends. The total number of received nominations was divided by the maximum number of possible nominations per group. Higher scores reflect greater network cohesion. Group density scores ranged from 0.73 to 12.55 (M\u0026thinsp;=\u0026thinsp;8.51, SD\u0026thinsp;=\u0026thinsp;1.82) and were standardized using z-scores.\u003c/p\u003e\u003cp\u003e\u003cem\u003eHierarchy.\u003c/em\u003e It represents the level of centrality within each group. It was measured as the average standard deviation of friendship nominations at the classroom level. Hierarchy scores ranged from 0.0 to 7.48, M\u0026thinsp;=\u0026thinsp;3.83 SD\u0026thinsp;=\u0026thinsp;0.95)\u003c/p\u003e\u003cp\u003e\u003cem\u003eGender ratio\u003c/em\u003e. It was calculated by dividing the number of girls by the number of boys in the classroom. Higher gender ratio values reflect a lower presence of opposite-gender classmates (ranged from 0.14 to 12.0, M\u0026thinsp;=\u0026thinsp;1.02, SD\u0026thinsp;=\u0026thinsp;0.64).\u003c/p\u003e\u003cp\u003e\u003cem\u003eClass size.\u003c/em\u003e It refers to the total number of students enrolled in a given classroom. In the context of the Spanish education system, where this study was conducted, national regulations stipulate a legal maximum of 30 students per classroom in compulsory education. Consequently, class sizes in this sample reflect variability within this normative threshold. In the present study, class size was measured as a continuous variable (ranged from 10 to 35, M\u0026thinsp;=\u0026thinsp;22.43, SD\u0026thinsp;=\u0026thinsp;4.78).\u003c/p\u003e\u003cp\u003e\u003cem\u003eEducational level.\u003c/em\u003e It was coded as a binary variable (0\u0026thinsp;=\u0026thinsp;primary education, 1\u0026thinsp;=\u0026thinsp;secondary education) to examine differences in classroom dynamics across educational stages.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Data Analysis\u003c/h2\u003e\u003cp\u003eA Latent Class Analysis (LCA) was conducted to identify distinct classroom profiles based on their structural and normative characteristics. Since the focus of the analysis is on the group dimension of the classroom, the data were aggregated at the classroom level. To ensure the validity of the indicators at the group level, only classrooms with 10 or more students were included. Six indicators were considered for the analysis; classroom level of peer victimization, classroom cohesion, hierarchy, gender ratio, classroom size (number of students) and aggressive class norm. All indicators were transformed into standardized scores (Z-scores) to ensure comparability and prevent scale differences from influencing model estimation. All analyses were conducted in Python using the following packages: scikit-learn [\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e] for Gaussian Mixture Modeling, pandas [\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e] for data manipulation, NumPy [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] for numerical computations, SciPy [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e] for statistical tests, and pingouin [\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e] for post-hoc analyses.\u003c/p\u003e\u003cp\u003eMultiple models with between two and five latent classes were estimated, and the selection of the optimal model was based on a combination of statistical fit and theoretical relevance criteria. The indices considered included the Akaike Information Criterion (AIC), the Bayesian Information Criterion (BIC), the adjusted BIC (aBIC), the log-likelihood, the entropy, as well as specific likelihood tests such as the Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) and the Bootstrapped Likelihood Ratio Test (BLRT). The estimation process aimed to identify the model that best represented the latent structure of classrooms in terms of social configuration, considering both statistical fit and parsimony and the theoretical interpretability of the number of classes. Given substantial heterogeneity among latent classes\u0026mdash;evidenced by significant differences in classroom characteristics (all F\u0026thinsp;\u0026gt;\u0026thinsp;14.74, p\u0026thinsp;\u0026lt;\u0026thinsp;.001), variable predictive capacity (R\u0026sup2; range\u0026thinsp;=\u0026thinsp;0.224), and heterogeneous regression coefficients (β range\u0026thinsp;\u0026gt;\u0026thinsp;0.38)\u0026mdash;a stratified analysis approach was adopted.\u003c/p\u003e\u003cp\u003eFor each latent class, multiple regression analyses examined victimization predictors (proportion of girls, class size, cohesion, hierarchy, and aggressive norms) using standardized variables. Heterogeneity was evaluated through ANOVA for mean differences, coefficient ranges across classes, and predictive capacity comparisons.\u003c/p\u003e\u003c/div\u003e"},{"header":"3. Results","content":"\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Correlations and descriptive analysis\u003c/h2\u003e\u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e presents the correlation matrix and descriptive statistics for all study variables. The analysis reveals several significant associations that provide insights into the interrelationships among classroom.\u003c/p\u003e\u003cp\u003eThe strongest correlation in the matrix was observed between group size and aggressive classroom norms (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;.573, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), indicating that smaller classrooms are significantly more likely to develop permissive attitudes toward aggression. This finding suggests an important structural constraint on normative development within classroom contexts. Cohesion and hierarchy demonstrated a substantial positive correlation (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.550, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), indicating that classrooms with stronger social bonds tend to have more pronounced status hierarchies. This relationship suggests that social organization and group structure are closely intertwined in classroom contexts.\u003c/p\u003e\u003cp\u003eGroup size showed additional positive correlations with both cohesion (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.313, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01) and hierarchy (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.342, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), suggesting that larger classrooms may facilitate both greater social connectedness and more differentiated status structures.\u003c/p\u003e\u003cp\u003eVictimization showed its strongest positive association with aggressive classroom norms (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.493, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), Victimization was also negatively correlated with group size (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;.378, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), suggesting that smaller classrooms are associated with higher levels of victimization. Additionally, victimization showed a positive correlation with cohesion (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.174, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01) and a negative correlation with girls' ratio (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;.114, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01).\u003c/p\u003e\u003cp\u003eGirls' ratio showed negative correlations with hierarchy (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;.109, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.01), indicating that female-dominated classrooms tend to have less pronounced status differentials. However, girls' ratio showed no significant association with aggressive classroom norms (\u003cem\u003er\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.004, \u003cem\u003ens\u003c/em\u003e), suggesting that gender composition influences social structure but not necessarily normative attitudes toward aggression.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eCorrelations and Descriptive Statistics of Study Variables\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"9\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u003cem\u003eM\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u003cem\u003eSD\u003c/em\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1. Victimization\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.01\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.04\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2. Cohesion\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e8.51\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1.82\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e.174**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3. Hierarchy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e3.83\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.95\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.045\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e.550**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4. Girls ratio\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.02\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.64\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.114**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.059\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.109**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5. Group Size\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e22.43\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e4.78\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.378**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e.313**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e.342**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.107**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e6. Aggressive class norm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.21\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.51\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e.493**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e-0.051\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.191**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e0.004\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026minus;\u0026thinsp;.573**\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003e\u0026mdash;\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e* Descriptive statistics show original scale values before standardization. Correlations were computed using standardized variables. M\u0026thinsp;=\u0026thinsp;Mean; SD\u0026thinsp;=\u0026thinsp;Standard Deviation. **\u003cem\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;.01.\u003c/em\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eModel Fit Indices for Latent Class Analysis\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"7\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eModel\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAIC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eBIC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eaBIC\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eEntropy\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eLMR-LRT (\u003cem\u003ep\u003c/em\u003e value)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eBLRT (\u003cem\u003ep\u003c/em\u003e value)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2-class\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e11195.176\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e11598.812\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e11138.415\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.887\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3-class\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10405.375\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e11013.097\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e10320.913\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.942\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4-class\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e10054.746\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e10866.554\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e9942.584\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.920\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.439\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5-class\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e8820.069\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e9835.963\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e8680.207\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.981\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.034\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.012\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"7\"\u003e\u003cem\u003eNote.\u003c/em\u003e AIC\u0026thinsp;=\u0026thinsp;Akaike Information Criterion; BIC\u0026thinsp;=\u0026thinsp;Bayesian Information Criterion; aBIC\u0026thinsp;=\u0026thinsp;sample-size adjusted Bayesian Information Criterion; LMR-LRT\u0026thinsp;=\u0026thinsp;Lo-Mendell-Rubin Likelihood Ratio Test; BLRT\u0026thinsp;=\u0026thinsp;Bootstrap Likelihood Ratio Test.\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec13\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Model selection\u003c/h2\u003e\u003cp\u003eBased on multiple fit indices, the 3-class model was selected as the optimal solution. The BIC showed its optimal point at 3 classes, with values increasing thereafter. The Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) was significant for 2-class and 3-class solutions (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001) but became non-significant for the 4-class model (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;=\u0026thinsp;.439), indicating that additional classes do not substantially improve fit. The 3-class model demonstrated excellent classification quality (entropy\u0026thinsp;=\u0026thinsp;0.942) and maintained interpretability while balancing statistical fit and parsimony. Although 4- and 5-class models showed lower AIC values, convergence issues and potential overfitting made the 3-class solution the most defensible choice. The comparative analysis of the estimated models (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e) shows substantial improvements in the fit indicators with an increasing number of classes, albeit with diminishing returns. Moving from 2 to 3 classes, a considerable reduction in the information criteria is observed, with a ΔAIC of 789.801, a ΔBIC of 585.715, and a ΔaBIC of 817.502, as well as an improvement in entropy (Δ\u0026thinsp;=\u0026thinsp;0.055). Furthermore, both the LMR-LRT and the BLRT are statistically significant, reinforcing the validity of the three-class model. The transition from 3 to 4 classes continues this trend with a ΔAIC of 350.629, a ΔBIC of 146.543, and a ΔaBIC of 378.329. However, a slight decrease in entropy is observed (Δ = -0.022), and the LMR-LRT is no longer significant, indicating that the improvement in fit does not justify the increase in model complexity (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). These results suggest that the three-class model represents an adequate balance between statistical fit and parsimony, offering an interpretable and robust solution for describing social heterogeneity between classrooms.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Latent Profile Analysis\u003c/h2\u003e\u003cp\u003eThree distinct latent classes emerged from the analysis, demonstrating significant heterogeneity in classroom social configurations (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Class 1 (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;80 classrooms, 11.6%) showed the lowest victimization levels of all classes (\u003cem\u003eM\u003c/em\u003e = -0.40, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.33) and below-average aggressive norms (\u003cem\u003eM\u003c/em\u003e = -0.12, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.33). This class is characterized by notably low cohesion (\u003cem\u003eM\u003c/em\u003e = -0.59, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.04) and below-average hierarchy (\u003cem\u003eM\u003c/em\u003e = -0.39, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.92). The most defining feature of this class is an extreme gender imbalance strongly favoring girls (girls ratio \u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.26, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.96), combined with below-average class size (\u003cem\u003eM\u003c/em\u003e = -0.70, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.78). These characteristics suggest an environment with minimal peer aggression but significant challenges in social organization and group cohesion.\u003c/p\u003e\u003cp\u003eClass 2 (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;130 classrooms, 18.9%) emerges as the most vulnerable and high-risk profile. It is characterized by the highest victimization levels across all classes (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.56, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.75), moderate cohesion (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.17, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.95), and below-average hierarchy (\u003cem\u003eM\u003c/em\u003e = -0.23, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;1.01). These classrooms are the smallest among all classes (\u003cem\u003eM\u003c/em\u003e = -0.98, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.51) and show a slight gender imbalance toward boys (girls ratio \u003cem\u003eM\u003c/em\u003e = -0.13, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.44). Class 2 exhibits the highest aggressive norms (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.38, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.36), creating a context of elevated social risk with substantial potential for peer conflict and bullying behaviors.\u003c/p\u003e\u003cp\u003eClass 3 (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;478 classrooms, 69.5%) represents the most normative and balanced profile among all identified classes. This class is characterized by slightly below-average victimization levels (\u003cem\u003eM\u003c/em\u003e = -0.09, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.37), moderate positive cohesion (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.19, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.80), and slightly above-average hierarchy (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.12, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.92). The gender distribution shows a slight imbalance toward boys (girls ratio \u003cem\u003eM\u003c/em\u003e = -0.15, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.38), while class size is considerably above the mean (\u003cem\u003eM\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.32, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.62). Aggressive norms are below average (\u003cem\u003eM\u003c/em\u003e = -0.16, \u003cem\u003eSD\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.18), indicating an environment with lower prevalence of aggressive behaviors and representing the most stable classroom climate.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\u003ch2\u003e3.4 Analysis of Variance and Effect Sizes\u003c/h2\u003e\u003cp\u003eAnalysis of variance revealed highly significant differences across all measured variables (\u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001), with effect sizes ranging from moderate to very large (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). Class size emerged as the most defining characteristic of the latent classes, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;276.81, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .447. Aggressive classroom norms showed comparable importance, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;250.35, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .422. These two dimensions collectively explained the majority of differences between latent classes. Victimization levels, representing the primary outcome variable, showed substantial between-class variation, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;136.20, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .284. Girls' ratio demonstrated strong differentiation between classes, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;119.69, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .259. Classroom cohesion and hierarchy showed smaller but significant contributions to class differentiation. Cohesion explained approximately 7.7% of between-class variance, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;28.60, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .077, while hierarchy contributed approximately 4.1% of variance, \u003cem\u003eF\u003c/em\u003e(2, 685)\u0026thinsp;=\u0026thinsp;14.74, \u003cem\u003ep\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;.001, \u003cem\u003eη\u0026sup2;\u003c/em\u003e = .041.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eDescriptive Statistics and Statistical Tests for Variables Across Latent Classes\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass 1 M (SD)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eClass 2 M (SD)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eClass 3 M (SD)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eF-Statistic\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003ePost Hoc Interpretation\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVictimization\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e-0.40 (0.33)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.56 (0.75)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-0.09 (0.37)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e136.20***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e2\u0026thinsp;\u0026gt;\u0026thinsp;3\u0026thinsp;\u0026gt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCohesion\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e-0.59 (1.04)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.17 (0.95)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.19 (0.80)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e28.60***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e2, 3\u0026thinsp;\u0026gt;\u0026thinsp;1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHierarchy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e-0.39 (0.92)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e-0.23 (1.01)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.12 (0.92)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e14.74***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3\u0026thinsp;\u0026gt;\u0026thinsp;1, 2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGirls ratio\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e1.26 (1.96)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e-0.13 (0.44)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-0.15 (0.38)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e119.69***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e1\u0026thinsp;\u0026gt;\u0026thinsp;2, 3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eClass Size\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e-0.70 (0.78)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e-0.98 (0.51)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.32 (0.62)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e276.81***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e3\u0026thinsp;\u0026gt;\u0026thinsp;1\u0026thinsp;\u0026gt;\u0026thinsp;2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAggressive class norm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e-0.12 (0.33)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.38 (0.36)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e-0.16 (0.18)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e250.35***\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e2\u0026thinsp;\u0026gt;\u0026thinsp;1, 3\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e*Note. M\u0026thinsp;=\u0026thinsp;Mean; SD\u0026thinsp;=\u0026thinsp;Standard deviation. Statistical significance was tested using one-way ANOVA. ***\u003cem\u003ep\u0026thinsp;\u0026lt;\u0026thinsp;.001. F values correspond to p\u0026thinsp;\u0026lt;\u0026thinsp;.001 for all variables.\u003c/em\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec16\" class=\"Section2\"\u003e\u003ch2\u003e3.5 Stratified Regression Analysis for Victimization\u003c/h2\u003e\u003cp\u003eRegression models revealed markedly different predictive capacity across latent classes (Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Class 2 showed the highest predictive capacity (\u003cem\u003eR\u003c/em\u003e\u0026sup2; = .354, adjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2; = .328), followed by Class 3 (\u003cem\u003eR\u003c/em\u003e\u0026sup2; = .200, adjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2; = .146) and Class 1 (\u003cem\u003eR\u003c/em\u003e\u0026sup2; = .130, adjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2; = .121). The \u003cem\u003eR\u003c/em\u003e\u0026sup2; range across classes (0.224) indicates substantial heterogeneity in victimization mechanisms. \u003cem\u003eClass 1\u003c/em\u003e (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;80), characterized by high proportion of girls and small classes, victimization was primarily predicted by internal classroom factors. Aggressive norms were the strongest predictor (β\u0026thinsp;=\u0026thinsp;.105, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .118), followed by hierarchy (β = \u0026minus;\u0026thinsp;.076, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .101). Crucially, class size showed no association with victimization (β\u0026thinsp;=\u0026thinsp;.001, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .034), suggesting that structural factors are less relevant than internal group dynamics in these classrooms. \u003cem\u003eClass 2\u003c/em\u003e, (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;130) showed the highest predictive capacity. Class size was the dominant predictor (β = \u0026minus;\u0026thinsp;.451, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .158), indicating that smaller classes are strongly associated with greater victimization. Paradoxically, cohesion also predicted greater victimization (β\u0026thinsp;=\u0026thinsp;.351, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .046), suggesting that group cohesion may facilitate collective victimization dynamics in these classrooms. Class size showed completely different effects across classes: null in Class 1 (β\u0026thinsp;=\u0026thinsp;.001), very strong and negative in Class 2 (β = \u0026minus;\u0026thinsp;.451), and weak in Class 3 (β = \u0026minus;\u0026thinsp;.077). This extreme heterogeneity demonstrates that relationships between classroom structure and victimization are not uniform, but fundamentally depend on the classroom's latent profile. \u003cem\u003eClass 3\u003c/em\u003e (n\u0026thinsp;=\u0026thinsp;478) with large classes and relatively low victimization, showed a limited predictive capacity. Cohesion was the strongest predictor (β\u0026thinsp;=\u0026thinsp;.099, \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = .030), though with weak effect.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eStratified Regression Analysis for Victimization by Latent Class\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass 1\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eClass 2\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eClass 3\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eβ (\u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eβ (\u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eβ (\u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eGirls ratio\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.010 (0.012)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.072 (0.012)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.015 (0.003)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eClass size\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.001 (0.034)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-0.451** (0.158)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.077 (0.043)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCohesion\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.035 (0.023)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.351** (0.046)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.099 (0.030)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHierarchy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-0.076 (0.101)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.018 (0.002)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-0.022 (0.000)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAggressive class norm\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e0.105* (0.118)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e0.011 (0.025)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e0.062 (0.052)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e.200\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e.354\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e.130\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eAdjusted \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e.146\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e.328\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e.121\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"4\"\u003e\u003cem\u003eNote.\u003c/em\u003e β\u0026thinsp;=\u0026thinsp;standardized regression coefficient; \u003cem\u003eR\u003c/em\u003e\u0026sup2;\u003csub\u003ei\u003c/sub\u003e = individual \u003cem\u003eR\u003c/em\u003e\u0026sup2; contribution. Class 1\u0026thinsp;=\u0026thinsp;Internal Conflict profile (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;80); Class 2\u0026thinsp;=\u0026thinsp;Structural Vulnerability profile (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;130); Class 3\u0026thinsp;=\u0026thinsp;Resilient profile (\u003cem\u003en\u003c/em\u003e\u0026thinsp;=\u0026thinsp;478).\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e|β| \u0026gt;0.1. ** |β| \u0026gt;0.2.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec17\" class=\"Section2\"\u003e\u003ch2\u003e3.6 Educational Levels and Profiles Distribution\u003c/h2\u003e\u003cp\u003eNext, we did an analysis of the three profiles across educational stages (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Analysis of class distribution reveals important developmental patterns in classroom social configurations. Class 3 (normative profile) appears more prevalent in secondary education with higher victimization in primary education. Class 2 (high-risk profile) shows a similar presence in both educational stages with higher victimization in primary education. Finally, class 1 (female-dominated profile) is predominant in secondary education with similar levels of victimization in both educational levels, potentially reflecting developmental differences in how gender composition affects classroom dynamics.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"4. Discussion","content":"\u003cp\u003eThe goal of this study was to identify distinct classroom social configurations based on group cohesion, status hierarchy, aggressive class norms, peer victimization, class size, and gender ratio. Using latent profile analysis, three distinct classroom profiles emerged. The emergence of three latent classes provides empirical support for the theoretical premise that classrooms function as complex social ecosystems with distinct organizational patterns [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Certain combinations of variables tend to appear together in consistent patterns, creating classrooms that are more or less conducive to bullying. The most predominant profile, representing nearly 70% of classrooms, suggests that most educational environments naturally develop protective characteristics against peer victimization. This finding aligns with W\u0026ouml;lfer and Scheithauer [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e], who noted that cohesion is generally associated with socioemotional adjustment and collaboration.\u003c/p\u003e\u003cp\u003eA more vulnerable profile presents a pattern, characterized by high cohesion, aggressive norms, and paradoxically small class size. The stratified regression analysis revealed that this profile showed the highest predictive capacity (R\u0026sup2; = .354), with class size emerging as the dominant predictor. This counterintuitive indicating that bullying can intensify when social networks are dense and opportunities to escape bullies' control are limited [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]. Particularly striking is the positive association between cohesion and victimization in these vulnerable contexts (β\u0026thinsp;=\u0026thinsp;.351), suggesting what Garandeau and Cillessen [\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e] described as \"false cohesion\", where group unity is achieved through shared targeting of victims rather than genuine social bonds. In these settings, bullies often occupy central positions within the classroom's social network, making it harder for other students to intervene [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. This parallels Rambaran et al. [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e], who warned that cohesion may become maladaptive when tied to aggressive group norms. In such contexts, bullies often occupy high-status positions [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], and peer reinforcement further stabilizes their dominance [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. This results are consistent with the evidence provided that in classrooms with high network density and strong embeddedness [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e] or hierarchy [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] victimized students are strongly disliked, highlighting how cohesion can magnify hierarchical inequalities.\u003c/p\u003e\u003cp\u003eRegarding the class size the vulnerable latent profile also illustrates this mechanism, combining high cohesion with small class size and aggressive norms. This combination magnifies power asymmetries and reduces victims' ability to escape peer control, supporting Garandeau et al. [\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e], who found that victimization intensifies when networks are dense. It also mirrors findings from Coelho and Rom\u0026atilde;o [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e] and Fekkes et al. [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], who showed that smaller classes can heighten bullying risks when peer structures are rigid and coercive leaders dominate.\u003c/p\u003e\u003cp\u003eThe least common profile (female-dominated profile) characterized by gender imbalance favoring girls yet showing the lowest victimization levels, and with the lowest levels of social structure (cohesion and hierarchy). This configuration demonstrates that victimization was primarily predicted by aggressive norms (β\u0026thinsp;=\u0026thinsp;.105) while class size showed no association (β\u0026thinsp;=\u0026thinsp;.001), suggesting that internal normative processes rather than structural factors determine outcomes. The high female presence may offer a buffering effect against victimization, possibly due to different socialization processes and interactional styles [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]. The notably low cohesion indicates that protection from victimization does not necessarily translate to connectedness in the classroom.\u003c/p\u003e\u003cdiv id=\"Sec19\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Implications for Educational Practice\u003c/h2\u003e\u003cp\u003eThese findings have significant implications for intervention. The variation in predictive patterns across profiles (R\u0026sup2; range\u0026thinsp;=\u0026thinsp;0.224) demonstrates that effective interventions must be tailored to specific classroom social configurations rather than applying universal approaches. For classrooms characterized by structural vulnerability, interventions should focus on restructuring social hierarchies and monitoring group dynamics. Special attention should be paid to the paradoxical role of cohesion in these contexts, ensuring that group unity is built around prosocial rather than exclusionary practices. Classrooms that demonstrate protective characteristics can benefit from approaches that reinforce and consolidate their existing strengths. Gender-imbalanced classrooms with low victimization require interventions that strengthen social cohesion while maintaining their naturally protective characteristics. Activities that promote inclusive group formation while leveraging the prosocial tendencies associated with higher female representation may be particularly effective.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec20\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Limitations and Future Directions\u003c/h2\u003e\u003cp\u003eSeveral limitations should be acknowledged. First, the cross-sectional design limits causal inferences about relationships between classroom characteristics and victimization outcomes. This prevents observation of how classroom dynamics evolve. Longitudinal studies are needed to establish whether classroom profiles represent stable configurations or dynamic states that change over time. Second, although the focus was on group characteristics, many bullying incidents relate to individual or family factors not considered in this study. Additionally, no information was collected on teacher behaviors or school-wide policies, both of which could influence classroom dynamics. Finally, although the sample is large and represents widely used programs in Spain, these profiles should be validated in other countries and educational contexts to assess their applicability across different environments\u003c/p\u003e\u003c/div\u003e"},{"header":"5. Conclusion","content":"\u003cp\u003eThis study demonstrates that classrooms are not uniform contexts for peer victimization but social systems with distinct organizational patterns. The identification of three distinct configurations provides a framework for understanding classroom-level heterogeneity in peer victimization processes and reinforces the consideration of bullying as a group-driven phenomenon embedded in peer dynamics and social norms. Such finding avoid the assumption that all classrooms require identical interventions.\u003c/p\u003e"},{"header":"Abbreviations","content":"\u003cp\u003eAIC - Akaike Information Criterion\u003cbr\u003e\u0026nbsp; aBIC - Sample-size adjusted Bayesian Information Criterion\u003cbr\u003e\u0026nbsp; ANOVA - Analysis of Variance\u003cbr\u003e\u0026nbsp; BIC - Bayesian Information Criterion\u003cbr\u003e\u0026nbsp; BLRT - Bootstrap Likelihood Ratio Test\u003cbr\u003e\u0026nbsp; GMM - Gaussian Mixture Model\u003cbr\u003e\u0026nbsp; LCA - Latent Class Analysis\u003cbr\u003e\u0026nbsp; LMR-LRT - Lo-Mendell-Rubin Likelihood Ratio Test\u003cbr\u003e\u0026nbsp; M - Mean\u003cbr\u003e\u0026nbsp; SD - Standard Deviation\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch3\u003e\u003cstrong\u003eEthics Approval and Consent to Participate\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThis study was approved by the Research Ethics Committee of Hospital General Universitario Gregorio Mara\u0026ntilde;\u0026oacute;n (Code: Sociescuela-GM, Protocol version 1.0, dated February 21, 2025). The committee issued a favorable opinion on March 3, 2025 (Act 03/2025), certifying that the study follows legally established requirements and is pertinent for implementation. The study was approved for conduct without consent from source subjects for data utilization, as the research involves secondary analysis of anonymized educational data collected through the Sociescuela program. The committee confirmed that the protocol meets necessary requirements of suitability in relation to study objectives and that foreseeable risks and inconveniences to subjects are justified. All procedures were performed in accordance with the ethical standards of the institutional research committee and with the 1964 Helsinki declaration and its later amendments.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eConsent for Publication\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eNot applicable. This study involved secondary analysis of anonymized data collected through the Sociescuela program. No individual participant data that could lead to identification is presented in this manuscript.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eAvailability of Data and Materials\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThe datasets analyzed during the current study are available from the corresponding author on reasonable request, subject to appropriate ethical approval and data sharing agreements. Raw data cannot be shared publicly due to privacy restrictions and ethical considerations related to student information, but aggregated and anonymized datasets supporting the conclusions of this article may be made available to qualified researchers upon request.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eCompeting Interests\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThe authors declare that they have no competing interests.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eThis research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors. The Sociescuela program data collection was supported by participating educational institutions.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eAuthors\u0026apos; Contributions\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eB.F.U. contributed to data collection, database processing, and manuscript writing. A.U. participated in database processing and manuscript writing. E.V.M. provided critical review and contributed to manuscript writing. V.S.L. designed the study, provided critical review and contributed to manuscript writing. J.M.B. (corresponding author) conducted data analysis, developed the methodology, supervised the study, and contributed to manuscript writing. All authors read and approved the final manuscript.\u003c/p\u003e\n\u003ch3\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/h3\u003e\n\u003cp\u003eWe acknowledge the participation of students, teachers, and educational institutions involved in the Sociescuela program. We thank the Research Ethics Committee of Hospital General Universitario Gregorio Mara\u0026ntilde;\u0026oacute;n for their ethical review and approval of this study. Special appreciation goes to all the schools and educational professionals who facilitated data collection and supported this research initiative. We also thank the students who participated in the sociometric assessments that made this research possible\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of generative AI and AI-assisted technologies in the writing process.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDuring the preparation of this work, the author(s) used DeepL Write and Grammarly to improve their English writing. After using this tool/service, the author(s) reviewed and edited the content as needed and took(s) full responsibility for the content of the published article.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n \u003cli\u003eHalliday S, Gregory T, Taylor A, Diggins E, Turnbull D. Adolescent mental health during the COVID-19 pandemic: A longitudinal study. J Affect Disord. 2021;290:336-45.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eUNESCO. Behind the numbers: Ending school violence and bullying. Paris: United Nations Educational, Scientific and Cultural Organization; 2019.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eOlweus D. Bullying at school: What we know and what we can do. Oxford: Blackwell Publishing; 1993.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSalmivalli C. Bullying and the peer group: A review. Aggression Violent Behav. 2010;15(2):112-20.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLaninga-Wijnen L, Steglich C, Harakeh Z, Vollebergh W, Veenstra R, Dijkstra JK. The role of prosocial and aggressive popularity norm combinations in prosocial and aggressive friendship processes. J Youth Adolesc. 2020;49(3):645-63.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eW\u0026ouml;lfer R, Scheithauer H. Social influence and bullying behavior: Intervention-based network dynamics of the fairplayer.manual bullying prevention program. Aggressive Behav. 2014;40(4):309-19.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eRambaran AJ, Hopmeyer A, Schwartz D, Steglich C, Badaly D, Veenstra R. Academic functioning and peer influences: A short-term longitudinal study of network-behavior dynamics in middle adolescence. Child Dev. 2020;88(2):523-43.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePouwels JL, Souren PM, Lansu TA, Cillessen AH. Stability of peer victimization: A meta-analysis of longitudinal research. Dev Rev. 2016;40:1-24.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eDijkstra JK, Lindenberg S, Veenstra R. Beyond the class norm: Bullying behavior of popular adolescents and its relation to peer acceptance and rejection. J Abnorm Child Psychol. 2008;36(8):1289-99.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLaninga-Wijnen L, Harakeh Z, Garandeau CF, Dijkstra JK, Veenstra R, Vollebergh WAM. Classroom popularity hierarchy predicts prosocial and aggressive popularity norms across the school year. Child Dev. 2019;90(5):e637-e655.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eAhn HJ, Garandeau CF, Rodkin PC. Effects of classroom embeddedness and density on the social status of aggressive and victimized children. J Early Adolesc. 2010;30(1):76-101.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMart\u0026iacute;n-Babarro J, D\u0026iacute;az-Aguado MJ, Mart\u0026iacute;nez-Arias R, Steglich C. Power structure in the peer group: The role of classroom cohesion and hierarchy in peer acceptance and rejection of victimized and aggressive students. J Early Adolesc. 2017;37(9):1197-220.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSaarento S, Salmivalli C. The role of classroom peer ecology and bystanders\u0026apos; responses in bullying. Child Dev Perspect. 2015;9(4):201-5.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eSaarento S, K\u0026auml;rn\u0026auml; A, Hodges EVE, Salmivalli C. Student-, classroom-, and school-level risk factors for victimization. J Sch Psychol. 2013;51(4):421-34.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eAguilar-Pardo D, Mart\u0026iacute;nez-Fern\u0026aacute;ndez B, Colmenares F, Mart\u0026iacute;n-Babarro J. Peer likeability and victimization in young adolescents: Moderating effects of descriptive and status group norms for aggression and prosocial behaviour and network density. Br J Educ Psychol. 2022;92(3):925-54.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePozzoli T, Gini G. Active defending and passive bystanding behavior in bullying: The role of personal characteristics and perceived peer pressure. J Abnorm Child Psychol. 2010;38(6):815-27.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eCoelho VA, Rom\u0026atilde;o AM. The relation between social anxiety, social withdrawal and (cyber)bullying roles: A multilevel analysis. Comput Human Behav. 2018;86:218-26.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eFekkes M, Pijpers FI, Fredriks AM, Vogels T, Verloove-Vanhorick SP. Do bullied children get ill, or do ill children get bullied? A prospective cohort study on the relationship between bullying and health-related symptoms. Pediatrics. 2006;117(5):1568-74.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eKowalski RM, Limber SP. Psychological, physical, and academic correlates of cyberbullying and traditional bullying. J Adolesc Health. 2013;53(1):S13-S20.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eCross D, Shaw T, Hadwen K, Cardoso P, Slee P, Roberts C, et al. Longitudinal impact of the Cyber Friendly Schools program on adolescents\u0026apos; cyberbullying behavior. Aggressive Behav. 2022;42(2):166-80.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eBusching R, Krah\u0026eacute; B. The girls set the tone: Gendered classroom norms and the development of aggression in adolescence. Pers Soc Psychol Bull. 2015;41(5):659-76.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLawson MA, Alameda-Lawson T, Downer J, Anderson E. Analyzing sub-population profiles and risk factors for school bullying. Child Youth Serv Rev. 2013;35(6):973-83.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePan Y, Liu H, Lau P, Luo F. A latent transition analysis of bullying and victimization in Chinese primary school students. PLoS One. 2017;12(8):e0182802.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eLiu J, Guo S, Weissman R, Liu H. Investigating factors associated with bullying utilizing latent class analysis among adolescents. Sch Psychol Int. 2021;42(1):11-32.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eJenkins LN, Snyder Kaminski S, Miller M. Bystander Intervention in Bullying: Differences Across Latent Profiles. Int J Bullying Prev. 2021;3(2):130-7.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWaasdorp TE, Bradshaw CP. Examining variation in adolescent bystanders\u0026apos; responses to bullying. School Psychol Rev. 2018;47(1):18-33.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eFerreira-Junior V, Valente JY, Sanchez ZM. Examining associations between race, gender, alcohol use, school performance, and patterns of bullying in the school context: A latent class analysis. J Interpers Violence. 2021;37(15-16):NP12857-NP12880.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePivec T, Horvat M, Ko\u0026scaron;ir K. Psychosocial characteristics of bullying participants: A person-oriented approach combining self- and peer-report measures. Child Youth Serv Rev. 2023;144:106729.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eGaffney H, Ttofi MM, Farrington DP. What works in anti-bullying programs? Analysis of effective intervention components. J Sch Psychol. 2021;85:37-56.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMart\u0026iacute;n-Babarro J. Assessment and detection of peer-bullying through analysis of the group context. Psicothema. 2014;26(3):357-63.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eWasserman S, Faust K. Social network analysis: Methods and applications. Cambridge: Cambridge University Press; 1994.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003ePedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine learning in Python. J Mach Learn Res. 2011;12:2825-30.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eMcKinney W. Data structures for statistical computing in Python. In: van der Walt S, Millman J, editors. Proceedings of the 9th Python in Science Conference; 2010. p. 51-6.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eHarris CR, Millman KJ, van der Walt SJ, Gommers R, Virtanen P, Cournapeau D, et al. Array programming with NumPy. Nature. 2020;585(7825):357-62.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eVirtanen P, Gommers R, Oliphant TE, Haberland M, Reddy T, Cournapeau D, et al. SciPy 1.0: Fundamental algorithms for scientific computing in Python. Nat Methods. 2020;17(3):261-72.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eVallat R. Pingouin: Statistics in Python. J Open Source Softw. 2018;3(31):1026.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eGarandeau CF, Lee IA, Salmivalli C. Inequality matters: classroom status hierarchy and adolescents\u0026apos; bullying. J Youth Adolesc. 2019;48(6):1212-28.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eGarandeau CF, Cillessen AHN. From indirect aggression to invisible aggression: A conceptual view on bullying and peer group manipulation. Aggression Violent Behav. 2006;11(6):612-25.\u0026nbsp;\u003c/li\u003e\n \u003cli\u003eCoelho VA, Sousa V. Differential effectiveness of a middle school social and emotional learning program: does setting matter? J Youth Adolesc. 2018;47(9):1978-1991.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Bullying, Peer victimization, Latent class analysis, Cohesion, Hierarchy, Classroom norms, Gender ratio, Class size","lastPublishedDoi":"10.21203/rs.3.rs-7678286/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7678286/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003e\u003cstrong\u003eBackground: \u003c/strong\u003eBullying is a complex group phenomenon, influenced by classroom social conditions that cannot be understood solely through individual characteristics. While previous research has examined factors like cohesion, hierarchy, norms, and structural characteristics separately as gender ratio and class size little is known about how these elements combine to form distinct social configurations associated with different bullying levels. This study aimed to identify latent classroom profiles based on multiple social and structural indicators and examine their relationship with peer victimization patterns.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMethods: \u003c/strong\u003eA latent class analysis was conducted with 19,708 students from 746 classrooms across primary and secondary education in Spain, participating in the Sociescuela program. Data were collected through computer-based sociometric peer nominations assessing victimization, cohesion, hierarchy, aggressive norms, class size, and gender composition. Classroom-level indicators were standardized and analyzed using Gaussian Mixture Models. Model selection was based on multiple fit indices including AIC, BIC, aBIC, entropy, and likelihood ratio tests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eResults: \u003c/strong\u003eA three-class solution provided the optimal balance between statistical fit and interpretability. A female-dominated low-risk profile (Class 1, n = 80 classrooms, 11.6%) showed the lowest victimization levels (M = -0.40), below-average aggressive norms (M = -0.12), low cohesion (M = -0.59) and gender imbalance favoring girls (M = 1.26). A high-risk vulnerable profile (Class 2, n = 130 classrooms, 18.9%) exhibited the highest victimization levels (M = 0.56), the highest aggressive norms (M = 0.38), the smallest class sizes (M = -0.98), and high cohesion (M = 0.17), representing the most problematic classroom environment. A normative balanced profile (Class 3, n = 478 classrooms, 69.5%) demonstrated slightly below-average victimization (M = -0.09), moderate positive cohesion (M = 0.19), larger class sizes (M = 0.32), and below-average aggressive norms (M = -0.16), representing the most stable and typical classroom climate.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConclusions: \u003c/strong\u003eThe study identified three distinct classroom social configurations with implications for bullying prevention. Findings emphasize that victimization risk depends on combinations of social and structural factors. Class 2 represents the highest-risk environment requiring intensive intervention, while Classes 1 and 3 show different protective mechanisms. Results support intervention strategies tailored to specific classroom profiles rather than applying universal approaches.\u003c/p\u003e","manuscriptTitle":"Mapping Classrooms and Peer Victimization Using a Latent Profile Approach: Network, Normative and Demographic Factors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-10 11:49:16","doi":"10.21203/rs.3.rs-7678286/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"e046675c-a525-4428-be5c-4b70c1fadb2a","owner":[],"postedDate":"November 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-01-27T11:13:19+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-10 11:49:16","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7678286","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7678286","identity":"rs-7678286","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}