Ecological and Predictive Indicators of Social and Emotional Skills among Korean Adolescents: A Person-Centered Analysis within the OECD SSES Framework

Ecological and Predictive Indicators of Social and Emotional Skills among Korean Adolescents: A Person-Centered Analysis within the OECD SSES Framework | 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 Ecological and Predictive Indicators of Social and Emotional Skills among Korean Adolescents: A Person-Centered Analysis within the OECD SSES Framework Nayoung Kim This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7891642/v1 This work is licensed under a CC BY 4.0 License Status: Published Journal Publication published 13 Mar, 2026 Read the published version in Child Indicators Research → Version 1 posted You are reading this latest preprint version Abstract This study investigates the multidimensional structure and ecological predictors of Korean adolescents’ socio-emotional skills using a person-centered, data-driven approach grounded in the OECD Social and Emotional Skills (SSES) framework. Utilizing nationally representative data from the Korean Children and Youth Panel Survey (KCYPS; N = 2,275), latent profile analysis identified three distinct subgroups: (1) Holistically Competent and Well-Adjusted (20.1%), (2) Emotionally Sensitive but Socially Withdrawn (47.5%), and (3) Cognitively Engaged but Emotionally Detached (32.4%). To determine ecological predictors of subgroup membership, a stacked ensemble machine learning model (CAWPE) was employed, integrating individual, familial, and school-level variables. The most influential predictors included parental cognitive empathy and creativity, adolescents’ self-esteem, peer and teacher relationships, household income, and engagement in leisure and career-related activities. Findings reveal that socio-emotional development in early adolescence is not a linear process but a multidimensional interplay of psychological, relational, and contextual factors within proximal ecological systems. By empirically operationalizing the OECD SSES framework using large-scale national data, this study provides a novel cross-level validation of socio-emotional skill indicators and demonstrates how person-centered and predictive analytics can inform evidence-based, developmentally targeted interventions and policy design. Social and emotional skills Child indicators OECD SSES framework Latent profile analysis Machine learning Figures Figure 1 1 Introduction Adolescents today are growing up amid rapid transformations in social structures and value systems. The aftermath of the COVID-19 pandemic, marked by prolonged social isolation and emotional instability, has posed serious challenges to their development and highlighted the urgent need for robust data to guide policy responses (Brown et al., 2021 ; Jones et al., 2022 ; Smith et al., 2021 ). Within this context, socio-emotional skills (SoES)—the abilities to regulate emotions, empathize with others, and make responsible decisions—have emerged not only as critical developmental assets but also as a crucial indicator for monitoring national child well-being (Blakemore & Mills, 2014 ; Casey, 2015 ). As foundational competencies for resilience and social adaptation, these indicators are essential for policymakers seeking to design effective support systems and evaluate the psychological health of the youth population. Understanding the structure and predictors of SoES is therefore a prerequisite for evidence-based adolescent policy and practice (Durlak et al., 2011 ). To conceptualize and promote these skills, various frameworks have been proposed. Among them, the CASEL (Collaborative for Academic, Social, and Emotional Learning) model has become a dominant paradigm, offering a systematic structure for implementing social and emotional learning (SEL) in schools. It comprises five core competencies—self-awareness, self-management, social awareness, relationship skills, and responsible decision-making (CASEL, 2015)—and has gained widespread adoption due to its strong curricular applicability. However, despite its practical utility, the CASEL model has been critiqued for its relatively static and linear conceptualization of learners’ internal competencies, which may overlook the contextual and dynamic nature of socio-emotional development. Empirical evidence supports this critique. For example, Ross et al. ( 2019 ) demonstrated through a large-scale longitudinal study that adolescents’ socio-emotional skills develop along diverse and non-linear trajectories that vary by skill domain and gender. Some adolescents may exhibit strong empathy yet struggle to maintain peer relationships, whereas others may be socially outgoing but less adept at emotional regulation. Such asynchronous developmental patterns cannot be fully captured through variable-centered approaches, which emphasize mean-level associations. Accordingly, scholars have highlighted the importance of person-centered approaches that uncover heterogeneity and within-person configurations of multiple competencies (Blakemore & Mills, 2014 ; Ross et al., 2019 ). The person-centered perspective offers both theoretical and practical advantages by focusing on how different competencies interact within individuals rather than examining them in isolation. Latent Profile Analysis (LPA), in particular, enables researchers to identify statistically distinct subgroups based on patterns of socio-emotional functioning and to examine how these profiles relate to developmental outcomes (Jung & Wickrama, 2008 ). For instance, Ağırkan and Haspolat (2025) applied LPA to classify high-school students’ socio-emotional competence profiles and reported meaningful differences in academic adjustment and psychological well-being across subgroups, thereby underscoring the multidimensional and differentiated nature of socio-emotional competence. Building on this person-centered foundation, the present study adopts the OECD Social and Emotional Skills (SSES) framework to explore the heterogeneity of SoES within a developmental-ecological context. In contrast to practitioner-focused models like CASEL, the OECD SSES provides a robust theoretical model for understanding socio-emotional development as a dynamic, contextually embedded process (Chernyshenko et al., 2018 ; OECD, 2015 ). This ecological emphasis is rooted in its integration of the Big Five personality theory with Bronfenbrenner’s ecological systems theory (1994), which posits that competencies are malleable resources shaped by continuous interactions across individual, family, and school systems (Lerner et al., 2018). A comprehensive understanding based on this framework, therefore, inherently demands analytical approaches that can capture both intra-individual heterogeneity (i.e., how skills combine within a person) and inter-contextual dynamics (i.e., how environmental factors predict these combinations). Guided by this perspective, the study operationalizes the SSES framework—which has demonstrated strong cross-cultural validity (Kim, H., et al., 2020 ; Kim, K., et al., 2020 ; Woo et al., 2022 )—to classify adolescents’ socio-emotional profiles and identify key ecological predictors. Previous Korean research has largely relied on variable-centered methods, offering limited insight into how multiple competencies interact within ecological contexts. To address this gap, the present study employs a hybrid design integrating Latent Profile Analysis (LPA) with a stacked ensemble machine learning model (CAWPE). This combination enables both robust classification of socio-emotional profiles and identification of influential ecological predictors. By merging developmental-ecological theory with predictive analytics, this study advances a more comprehensive understanding of the dynamics underlying adolescent SoES. Ultimately, this research provides a methodological foundation to move beyond one-size-fits-all interventions, enabling the design of evidence-based, targeted policies aimed at the specific needs of distinct adolescent subgroups. The study addresses the following research questions: What latent profiles characterize adolescents’ socio-emotional skills? Which ecological factors at the individual, family, and school levels significantly predict membership in each profile? 2 Literature Review 2.1 Conceptualizing Socio-Emotional Skills Socio-emotional skills (SoES) represent essential developmental assets that foster adolescents’ psychological well-being, interpersonal competence, and academic adjustment. These skills encompass a broad spectrum of social and emotional abilities—including emotional regulation, empathy, collaboration, and responsible decision-making—that enable adaptive functioning within social contexts (Blakemore & Mills, 2014 ; Durlak et al., 2011 ). Rather than being innate traits, SoES are best understood as developmental capacities that emerge through ongoing interactions between individual dispositions (e.g., temperament, personality) and environmental influences (OECD, 2015 ). Thus, socio-emotional skills are dynamic, context-sensitive competencies that evolve across time and experience. The OECD Social and Emotional Skills Framework (SSES) offers a comprehensive theoretical model for conceptualizing this developmental process. Grounded in the Big Five personality structure—openness, conscientiousness, extraversion, agreeableness, and emotional stability—the SSES reconceptualizes personality traits as malleable socio-emotional competencies (Chernyshenko et al., 2018 ). It categorizes these competencies into five broad domains: task performance, emotional regulation, collaboration, open-mindedness, and engagement with others. Each domain comprises specific subskills such as perseverance, optimism, empathy, and sociability, which have been empirically linked to adolescents’ academic success, psychological resilience, and social adjustment. A distinctive strength of the SSES lies in its dual perspective: it views socio-emotional skills as both personality-based psychological resources and plastic, context-responsive competencies. This approach contrasts with the widely adopted CASEL (Collaborative for Academic, Social, and Emotional Learning) model, which defines five core competencies—self-awareness, self-management, social awareness, relationship skills, and responsible decision-making (CASEL, 2015). While the CASEL model provides a practical framework for implementing school-based SEL programs, it tends to conceptualize these competencies as static functional domains. Recent systematic reviews of SEL measurement instruments (Martinez-Yarza, et al., 2023 ) have highlighted that the majority of existing tools are grounded in the CASEL framework, with limited attention to alternative conceptualizations such as the OECD SSES. This imbalance underscores the need for frameworks that better capture the multidimensional and context-sensitive nature of socio-emotional skills. The SSES, in contrast, expands this understanding by embedding socio-emotional skills within a developmental–ecological perspective, emphasizing within-person heterogeneity and cross-contextual interdependence (Lerner et al., 2018; OECD, 2015 ). To address the limitations inherent in the traditional variable-centered approach, recent scholarship has increasingly adopted a person-centered perspective in examining adolescents’ socio-emotional skills (SoES). In particular, research employing Latent Profile Analysis (LPA) has empirically demonstrated that adolescent SoES are not monolithic, but instead are expressed through qualitatively distinct and heterogeneous subgroups (e.g., Peetz et al., 2025 ). Nevertheless, current LPA-based studies exhibit several notable constraints: (a) they are primarily grounded in the CASEL theoretical framework; (b) they focus predominantly on Western populations, limiting cultural generalizability; and (c) they have seldom utilized integrative frameworks like the OECD’s Study on Social and Emotional Skills (SSES) to explore the multidimensional nature of socio-emotional profiles in conjunction with their broader ecological contexts. The SSES framework aligns closely with Bronfenbrenner’s ( 1994 ) ecological systems theory, which posits that development arises from reciprocal interactions between individuals and nested environmental systems—such as family, peers, schools, and communities. Extending this ecological foundation, the SSES conceptualizes socio-emotional competencies as multi-layered and adaptive resources shaped by continuous exchanges across ecological levels. In this regard, the SSES functions not merely as a personality model but as an applied developmental framework that integrates insights from developmental psychology and ecology to explain the dynamic and interactive nature of socio-emotional growth. In Korea, the increasing policy emphasis on social and emotional learning (SEL) has led to the integration of SEL components into the national curriculum and to growing interest in developing culturally grounded measurement instruments. Recent studies by Kim K., et al. ( 2020 ) and Woo et al. ( 2022 ) have validated the structural and cultural validity of the OECD SSES framework, confirming its applicability for capturing the socio-emotional characteristics and developmental diversity of Korean adolescents. These findings demonstrate that the SSES serves not only as a measurement tool but also as a conceptual and analytical lens for examining socio-emotional development across cultural contexts. Building upon this theoretical and empirical foundation, the present study extends the application of the SSES framework to a non-Western OECD context, empirically operationalizing its core indicators within nationally representative data. By integrating person-centered analysis (Latent Profile Analysis) with ecological and predictive modeling, this study advances the measurement and understanding of adolescent socio-emotional skills as multidimensional, dynamic, and contextually embedded phenomena—thereby addressing a key gap in comparative child indicator research. 2.2 Multilevel Predictors of Socio-Emotional Skills Adolescents’ socio-emotional skills (SoES) are shaped not by a single factor but through a multilevel ecological developmental process in which individual, familial, school, and community systems interact dynamically (Bronfenbrenner, 1994 ). Within this framework, the development of SoES is driven primarily by proximal processes—the everyday relational interactions with parents, teachers, and peers—rather than by distal structural or institutional influences. Integrating the OECD Social and Emotional Skills (SSES) framework with Bronfenbrenner’s ecological model, this study conceptualizes SoES as ecologically embedded and measurable developmental indicators, rather than as fixed psychological traits. Individual-Level Factors At the individual level, psychological resources such as self-esteem, emotional stability, aggression, and impulsivity serve as foundational predictors of socio-emotional competence (Burke & Loeber, 2016 ; Orth et al., 2012 ). High self-esteem and emotional stability foster empathy, prosocial behavior, and resilience, whereas aggression and impulsivity undermine emotional regulation and interpersonal adjustment. In addition, lifestyle and self-regulatory behaviors—including adequate sleep, regular physical activity, and balanced media use—operate as moderators that promote emotional well-being and facilitate positive peer relationships (Portela-Pino et al., 2021 ; Yu & Lee, 2023 ). Academic variables also play a key role: academic achievement and engagement strengthen motivation and self-regulation, while academic stress and burnout have been linked to decreased socio-emotional functioning (Salmela-Aro & Upadyaya, 2020 ). Collectively, these individual-level indicators align with the task performance and emotional regulation domains of the SSES framework. Family-Level Factors The family constitutes the primary context for emotional socialization during adolescence. Parental emotional support—characterized by warmth, empathy, and responsiveness—directly enhances adolescents’ emotional regulation and interpersonal trust (Datu & Restubog, 2020 ; Ong et al., 2018). Beyond relational dynamics, structural family resources such as parental education and household income provide cognitive and affective stimulation that fosters socio-emotional growth (Thompson et al., 2013 ). Parental cognitive empathy and creativity, in particular, contribute to adolescents’ open-mindedness and problem-solving capacities by modeling reflective and flexible thinking. Moreover, career-related parent–child interactions, including shared goal setting and discussions of future plans, strengthen autonomy and future orientation—skills conceptually situated within the task performance and open-mindedness domains of the SSES (Huhtala et al., 2014 ). Thus, family-level factors jointly represent both structural resources and interactive mechanisms that cultivate adolescents’ socio-emotional competencies. School-Level Factors The school serves as a core microsystem for socialization, where adolescents internalize social norms and interactional skills through daily engagement with teachers and peers. The quality of teacher–student relationships is consistently associated with school belonging, learning motivation, and emotional regulation (Roorda et al., 2011 ). Similarly, peer acceptance and supportive friendships act as protective buffers that mitigate externalizing behaviors and social withdrawal (Glick & Rose, 2011 ). Participation in extracurricular and self-directed activities—such as clubs, mentoring programs, and career exploration—further reinforces cooperation, empathy, and self-control, contributing to both academic achievement and long-term social responsibility (Durlak et al., 2011 ; Han et al., 2014 ; Sklad et al., 2012 ). Furthermore, experiences of career counseling and identity exploration have been identified as salient predictors of socio-emotional growth, promoting emotional maturity and goal-directed self-regulation (Savickas, 2013 ; Napolitano et al., 2021 ). Collectively, these school-level indicators correspond to the collaboration and engagement with others domains of the SSES, underscoring the school’s role as an everyday ecological context for the development of emotional and interpersonal competencies. In conclusion, existing research offers compelling evidence that factors at the individual, familial, and school levels are significant ecological predictors of students' sense of emotional and social well-being (SoES). Nevertheless, much of this work has predominantly employed traditional regression methods that examine the linear and independent contributions of each predictor. Such analytical approaches are often inadequate for capturing the intricate, non-linear, and interactive dynamics characteristic of complex ecological systems. As a result, they fall short in addressing a fundamental policy-relevant question: which among these numerous factors exert the greatest influence on adolescent developmental outcomes? This methodological limitation underscores the need for more sophisticated analytical techniques. The current study responds to this gap by adopting a predictive analytics framework, specifically machine learning, to not only corroborate established ecological associations but also to identify the most influential indicators of subgroup membership. This approach advances beyond mere correlational analysis, providing a more robust and actionable evidence base to inform targeted policy interventions. 3 Methods 3.1 Data and Participants This study drew on data from the fourth wave (2021) of the Korean Children and Youth Panel Survey (KCYPS) 2018, conducted by the National Youth Policy Institute (NYPI) of Korea. The KCYPS is a nationally representative longitudinal study designed to examine developmental, psychosocial, and educational indicators across key life transitions in childhood and adolescence (Ha et al., 2017 ). The survey includes two cohorts—one initiated in the fourth grade of elementary school and another in the first grade of middle school—and provides rich information on adolescents’ socio-emotional, behavioral, and contextual characteristics. The present study focused on the elementary school cohort, who were in their first year of middle school (Grade 7) at the time of the fourth wave. This developmental stage represents early adolescence, a critical transition period during which socio-emotional skills and self-regulatory competencies undergo rapid change. While the KCYPS has a longitudinal design, the current analysis used cross-sectional data from Wave 4, as the socio-emotional skills (SoES) indicators were uniquely collected in a special module during this wave. Consequently, longitudinal analyses (e.g., intra-individual growth trajectories) were not feasible. Nonetheless, the present design offers a valuable baseline validation of socio-emotional skill indicators within a nationally representative dataset, serving as an empirical foundation for future longitudinal research. The initial sample comprised 2,607 adolescents. After excluding participants with missing values on core SoES indicators, the final analytic sample consisted of 2,275 students (50.3% male, n = 1,145; 49.7% female, n = 1,130). Attrition analysis indicated no significant group differences in demographic and ecological variables between retained and excluded participants, suggesting that the final sample remained demographically representative of the original cohort. 3.2 Measures Social and Emotional Skills Adolescents’ socio-emotional skills (SoES) were defined and measured in accordance with the OECD Social and Emotional Skills Framework (SSES). The SSES conceptualizes socio-emotional skills as psychological resources that develop through continuous interaction between personality-based dispositions and environmental factors, encompassing five core domains: (1) task performance, (2) emotional regulation, (3) collaboration, (4) open-mindedness, and (5) engagement with others (Chernyshenko et al., 2018 ; OECD, 2015 ). While the OECD SSES covers a broad and multidimensional construct space, it is methodologically impractical to include all subdomains within a single survey . Therefore, this study operationalized the SSES framework by selecting theoretically representative and empirically validated indicators available in the Korean Children and Youth Panel Survey (KCYPS). Each selected variable was mapped to the SSES domains based on prior validation studies (Kim, H., et al., 2020 ; Woo et al., 2022 ) and its conceptual correspondence to the five core areas of socio-emotional competence. This process ensured that the operationalization achieved a representative balance across domains, maintaining the structural integrity and conceptual breadth of the SSES rather than focusing narrowly on any single subfactor. Specifically, indicators included perseverance (4 items, Cronbach's α = .610) for task performance, optimism (5 items, α = .822) for emotional regulation, cognitive empathy (as assessed by the Reading the Mind in the Eyes Test) for collaboration (28 items, α = .724), creative personality traits (30 items, α = .707) for open-mindedness, and sociability (5 items, α = .887) for engagement with others. All items were standardized (z-scores) prior to analysis to enable direct comparison across measurement scales and domains. The subdomains, representative indicators, and reliability coefficients (Cronbach’s α) used in this study are summarized in Table 1 . Table 1 Measures of socio-emotional skills (SSES Framework) Domain Representative Subskills Example Item Cronbach’s α (Number of items) Task Performance Perseverance “I don’t give up easily even when things are difficult.” 0.610 (4) Emotional Regulation Optimism “I tend to think positively about the future.” 0.822(5) Collaboration Cognitive empathy Score converted from responses to 28 items in the Reading the Mind in the Eyes Test (RMET), representing emotional recognition ability. 0.724(28) Open-Mindedness Creativity “I tend to be creative in my thinking and problem-solving.” 0.707(30) Engagement with Others Sociability “I find it easy to get along with others.” 0.887(5) Predictor Variables To identify the multilevel ecological predictors of adolescents’ socio-emotional competence, this study organized the independent variables within a theoretically grounded framework. The selection of predictors was informed by the OECD Social and Emotional Skills (SSES) framework, which integrates the Big Five personality model with Bronfenbrenner’s ( 1994 ) ecological systems theory... (Chernyshenko et al., 2018 ; Lerner et al., 2018). The inclusion of a comprehensive set of variables spanning these ecological domains was an intentional methodological strategy. This approach was designed to leverage the primary strength of machine learning: its ability to sift through a high-dimensional and complex feature space and identify the most salient predictive indicators in a data-driven manner, moving beyond the limitations of traditional regression models. Consistent with this dual theoretical and analytical strategy, the predictors were systematically categorized across three ecological levels: individual, family, and school. The specific variables included within each of these levels are described in detail below. Individual-Level Factors At the individual level, predictors were organized into four categories. The first category, psychological attributes, included self-esteem, emotional stability, aggression, attention, happiness, and depressive symptoms. The second category, behavioral and lifestyle variables, comprised average sleep duration, physical activity frequency, smartphone usage, and leisure time. The third category, academic and career-related factors, encompassed academic achievement, academic engagement, academic helplessness, career adaptability, and family discussions about career planning. Finally, delinquent behavior indicators, such as experiences of offline or online delinquency, were included. Family-Level Factors At the family level, predictors were grouped into three categories. The first, socio-demographic characteristics, included parental education, household income, and weekly working hours. The second, parental psychological resources, encompassed parents' cognitive empathy and creativity scores. The third category, parenting and relational dynamics, consisted of the frequency of parent-child communication, parental satisfaction with the child's academic progress, and parents' future educational expectations for their child. School-Level Factors At the At the school level, predictors were divided into two categories. The first, structural conditions, included school location (urban/rural) and the average number of instructional days per week. The second category, relational and psychological factors, comprised peer relationship quality, teacher-student relationship quality, and school satisfaction. A complete list of all predictive variables used in the machine learning model, along with their operational definitions and measurement scales, is summarized in Table 2 . Table 2 Predictive variables used in the classification of latent profiles Domain Predictive Variables (Number of items) Students Individual Backgrounds and Living Conditions Gender, sleep duration, leisure time, duration of physical activity, smartphone usage, smartphone dependency, annual participation frequency in various activities Physical and Psychological Characteristics Physical symptoms, health status, attention, aggression, happiness, depression, self-esteem, and cooperativeness Learning Academic achievement, academic engagement, academic helplessness, study time, and participation in private tutoring Career Career adaptability, frequency of career-related conversations, and decision-making status regarding future occupation Delinquent Behavior Presence and frequency of real-world delinquent experiences, cyber-delinquent experiences Home Home Background Parents’ highest level of education, monthly household income, economic status of the family, parents’ weekly working hours Parental Competencies parents’ cognitive empathy score, and parental creativity index Parent-Child Relationship Parental attitudes, time spent with parents, and time spent in conversations with parents Parents’ Educational Planning and Satisfaction with School Children’s satisfaction with school life, desired level of education for the future School School Characteristics Region, average number of school days per week Relationship and School Life Peer relationships, teacher relationships, and satisfaction with school life Note. Most continuous variables were measured on a 5-point Likert scale, while items related to competencies were measured on a 7-point scale 3.3 Analytical Strategy This study adopted a person-centered, two-stage analytical design to identify heterogeneous profiles of adolescents’ socio-emotional competencies and to determine their multilevel ecological predictors. The analytic framework combined latent profile analysis (LPA) for classification and ensemble-based machine learning for predictive modeling, thereby linking theory-driven typology with data-driven inference. In the first stage, LPA was employed to classify participants into latent subgroups based on five standardized socio-emotional indicators: creativity, perseverance, optimism, sociability, and empathy. LPA is a probabilistic clustering technique that identifies unobserved latent classes within a population and is particularly suited to detecting individual-level heterogeneity in psychological constructs. Model selection was guided by multiple fit indices, including the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), sample-size adjusted BIC (SA-BIC), entropy, and the Lo–Mendell–Rubin likelihood ratio test (LMR-LRT). A three-class solution was selected based on both statistical fit and theoretical interpretability, reflecting (1) Holistically Competent and Well-Adjusted, (2) Emotionally Sensitive but Socially Withdrawn, and (3) Cognitively Engaged but Emotionally Detached subgroups (see Tables 2 and 3 ). In the second stage, a supervised machine learning approach was used to examine the relative influence of individual, family, and school-level predictors on latent profile membership. Specifically, the Cross-validation Accuracy Weighted Probabilistic Ensemble (CAWPE) algorithm was applied (Large et al., 2019). CAWPE integrates predictions from multiple base classifiers—Random Forest, Gradient Boosting, Logistic Regression, Support Vector Machine (SVC), K-Nearest Neighbors, XGBoost, LightGBM, CatBoost, and Extra Trees—and assigns optimal weights according to cross-validated accuracy scores. This ensemble framework was chosen for its robustness against overfitting and its ability to model complex nonlinear interactions across ecological variables. Prior to modeling, data preprocessing included missing data treatment, outlier detection, and dimensionality reduction. Variables with more than 20% missingness were excluded. For the remaining variables, missing values were imputed using proximity-based imputation—the mean for continuous variables and the mode for categorical variables, based on the most similar observations (Breiman, 2001 ). Although more computationally intensive than simple imputation, this method effectively preserves nonlinear associations among predictors (Stekhoven & Bühlmann, 2012 ). To mitigate multicollinearity and computational burden, Principal Component Analysis (PCA) was performed for dimensionality reduction. The dataset was then randomly partitioned into training (70%) and testing (30%) subsets. To address class imbalance across latent profiles, the Synthetic Minority Oversampling Technique (SMOTE) was applied to the training data (Mishra & Singh, 2021 ). Model performance was evaluated using the Area Under the Receiver Operating Characteristic Curve (AUC) metric. The final CAWPE model achieved an AUC of 0.8788, indicating high predictive accuracy. Finally, partial dependence plots (PDPs) were generated to visualize the marginal effects of the top 20 most influential predictors on latent class membership, enhancing the interpretability of nonlinear relationships among ecological factors. 4 Results 4.1 Classification of Latent Profiles of Adolescents’ Socio-Emotional Competence Latent Profile Analysis (LPA) was conducted to identify distinct socio-emotional profiles among adolescents using five standardized indicators—creativity, perseverance, optimism, sociability, and empathy—derived from the OECD SSES framework. Model fit indices for the two- to six-profile solutions are summarized in Table 3 . Although the overall fit improved with additional classes (as indicated by reductions in AIC, BIC, and SA-BIC values), the rate of improvement diminished after the three-profile solution (Jedidi et al., 1997 ). Although the entropy value for the three-profile solution (0.748) fell slightly below the conventional threshold of 0.80 typically used to indicate strong classification quality (Clark & Muthén, 2009), it was deemed acceptable in light of the model's clear substantive interpretability. Moreover, while the Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) supported a five-profile solution (p < .001 for k = 5), solutions with four or more profiles produced classes with very small subgroup sizes (i.e., less than 5% of the sample), thereby limiting both interpretive stability and practical applicability (Jung & Wickrama, 2008 ). Consequently, the three-profile model was retained as the most parsimonious and theoretically coherent representation of the data. Figure 1 . presents the standardized mean scores across the five indicators for each latent group. The first group (32.4%, n = 736), labeled ‘Cognitively Engaged but Emotionally Detached’, exhibited moderate levels of cognitive and behavioral engagement but notably low cognitive empathy (Z = − 1.25), suggesting strong cognitive focus yet limited emotional responsiveness. The second group (47.5%, n = 1,081), labeled ‘Emotionally Sensitive but Socially Withdrawn’, showed relatively high empathy (Z > 0.6) but lower sociability (Z = − 0.4), reflecting emotional receptiveness combined with introverted and avoidant social tendencies. The third group (20.1%, n = 458), labeled ‘Holistically Competent and Well-Adjusted’, exhibited consistently high scores (Z > 0.5) across all cognitive, emotional, and social domains, representing a balanced and well-integrated profile of competence. The results reveal substantial heterogeneity in adolescents’ socio-emotional development. Notably, discrepancies between emotional understanding (empathy) and behavioral enactment (sociability and perseverance) across groups indicate that empathic awareness does not necessarily translate into social adaptability or proactive engagement. This finding underscores the importance of differentiated intervention strategies tailored to each subgroup’s specific strengths and vulnerabilities. Although cognitive empathy played a pivotal role in distinguishing between the profiles, classification was not determined by this indicator alone. As shown in Fig. 1 , meaningful differentiation also arose from variation in creativity, perseverance, and sociability. For example, the ‘Holistically Competent and Well-Adjusted’ group maintained uniformly high scores across all five indicators, whereas the ‘Emotionally Sensitive but Socially Withdrawn’ Group displayed a marked imbalance between affective and behavioral domains. While the single-item nature of the empathy measure warrants interpretive caution, the classification patterns were multidimensionally informed, not dependent on a single construct. Furthermore, variability in sociability (range = − 0.4 to + 0.63) and perseverance (range = − 0.43 to + 0.73) contributed meaningfully to subgroup differentiation, reinforcing that socio-emotional competence reflects a complex interplay of emotional, cognitive, and behavioral components rather than a linear continuum of skill development. Table 3 Fit indices and classification accuracy for latent profile models (2–6 Classes) Criterion Number of Latent Profiles 2 classes 3 classes 4 classes 5 classes 6 classes Information Indices AIC 31529.214 31007.603 30805.841 30686.937 30637.914 BIC 31620.890 31133.657 30966.274 30881.748 30867.103 SA-BIC 31570.055 31063.759 30877.313 30773.724 30740.016 Quality of Classification Entropy 0.881 0.748 0.754 0.745 0.753 Model Comparison Adjusted LMR (p) 0.001 < .001 < .001 < .001 0.079 B-LRT(p) 0.001 < .001 < .001 < .001 0.083 Classification Rate (Number, %) Class1 724(31.8) 736(32.4) 719(31.6) 617(27.1) 420 (18.5) Class2 1,551(68.2) 1,081(47.5) 918(40.4) 458(20.1) 595 (26.2) Class3 458(20.1) 527(23.2) 138(6.1) 963(42.3) Class4 111(4.9) 948(41.7) 156(6.9) Class5 Class6 114(5.0) 48 (2.1) 93 (4.1) Note While empathy showed the largest difference between profiles, meaningful variation was also observed in sociability (range: − 0.4 to + 0.63), perseverance (–0.43 to + 0.73), and creativity (–0.11 to + 0.78), indicating that classification was multidimensional. 4.2 Predictive Factors of Socio-Emotional Competence Subgroup Classification The Cross-validation Accuracy Weighted Probabilistic Ensemble (CAWPE) model achieved strong predictive performance (AUC = 0.8788), confirming the robustness of ecological predictors. Table 4 lists the twenty most influential variables, and Fig. 2 . illustrates the marginal effects of the major predictors on subgroup membership. Individual-Level Factors Behavioral and psychological characteristics were key differentiators among socio-emotional profiles. Adequate sleep (SL, SL1) and balanced physical activity (YTIM1J01w4) increased the likelihood of belonging to the ‘Holistically Competent and Well-Adjusted’ group, while excessive gaming (YTIM1L02w4) and unstructured smartphone use (YTIM1K01w4) predicted membership in the ‘Cognitively Engaged but Emotionally Detached’ group. Notably, purposeful digital engagement, such as information seeking (YMDA1B11w4), rather than mere usage volume, was more beneficial. Psychologically, adolescents with a strong sense of personal strengths (“I have many strengths,” YPSY3A03w4) were more likely to belong to the ‘Holistically Competent’ group, whereas those reporting low self-affirmation (“I have few things to be proud of,” YPSY3A05w4) tended to cluster in the ‘Emotionally Sensitive but Socially Withdrawn’ group. Family-Level Factors Family-level factors emerged as the most critical predictors, with parental cognitive empathy (PRME1Ascw4) identified as the single most dominant factor by a substantial margin. Its importance index (69.70) was more than three times higher than the second-ranked predictor (Time spent with parents on weekdays, 18.17), suggesting that a parent's capacity for empathy is the primary ecological foundation shaping adolescent socio-emotional profiles. Higher parental empathy substantially increased the likelihood of adolescents belonging to the ‘Holistically Competent and Well-Adjusted’ group while reducing the likelihood of belonging to the ‘Cognitively Engaged but Emotionally Detached’ group. Other family-level variables also showed significant predictive power. Regular parent–child interaction, especially on weekends (SPEND2, Importance = 17.71), increased the probability of ‘Holistic Competence’. Similarly, parental creative disposition (PPSY4Ascw4, Importance = 11.18) and higher household income (PINCOMEw4, Importance = 12.69) were important predictors that helped differentiate the ‘Holistically Competent’ group from the two less-adjusted profiles. School-Level Factors School-related predictors also significantly distinguished the groups. Positive teacher recognition (e.g., “The teacher considers me to be a smart student,” YEDU3A14w4) increased the probability of belonging to the 'Holistically Competent' group. A moderate willingness to seek teacher counseling (YEDU3A01w4) was optimal, differentiating balanced students from those overly dependent (Emotionally Sensitive) or emotionally distant (Emotionally Detached). Similarly, the frequency of conversations with a career counselor (YFUR2A06w4) showed a U-shaped relationship, with both very low and very high counseling frequencies linked to higher competence. Moderate peer interaction—during weekdays (YTIM1N02w4) and weekends (YTIM1N01w4) —was associated with 'Holistic Competence', while too little interaction aligned with 'Social Withdrawal' and excessive interaction with 'Emotional Detachment'. Moreover, higher agreement with “I do not want to be close to other students” (YEDU2A12w4) predicted the two less adjusted profiles. Table 4 Key predictors of adolescents’ socio-emotional competence classification Rank Variable Name Domain Questionnaire Importance Index 1 PRME1Ascw4 Parental Competencies Parent’s cognitive empathy score 69.70 2 SPEND1 Parent-Child Relationship Time spent with parents on weekdays 18.17 3 SPEND2 Parent-Child Relationship Time spent with parents on weekends 17.71 4 SL1 Individual Backgrounds and Living Conditions Sleep duration on weekends 14.50 5 YTIM1L02w4 Individual Backgrounds and Living Conditions Time spent playing with a computer on weekends 14.17 6 YTIM1N01w4 Relationship and School Life Time spent playing with friends outside of weekdays 13.81 7 YPSY3A03w4 Physical and Psychological Characteristics I perceive myself as having many strengths. 13.46 8 YFUR2A06w4 Relationship and School Life Frequency of career-related discussions with a school counselor 13.05 9 YTIM1M01w4 Individual Backgrounds and Living Conditions Time spent watching TV and playing on weekdays 12.97 10 SL Individual Backgrounds and Living Conditions sleep duration on weekdays 12.80 11 PINCOMEw4 Home Background Monthly average household income 12.69 12 YTIM1N02w4 Relationship and School Life Time spent with friends outside of weekends 12.16 13 YEDU2A12w4 Relationship and School Life I have no intention of becoming close with other children. 11.60 14 YPSY3A05w4 Physical and Psychological Characteristics I feel that I have few things to be proud of. 11.50 15 YMDA1B11w4 Individual Backgrounds and Living Conditions Frequency of smartphone use by purpose: view document 11.48 16 YTIM1K01w4 Individual Backgrounds and Living Conditions Time spent using a smartphone on weekdays 11.41 17 YEDU3A14w4 Relationship and School Life The teacher considers me to be a smart student. 11.33 18 YTIM1J01w4 Individual Backgrounds and Living Conditions Time spent on exercise and physical activity on weekdays 11.31 19 PPSY4Ascw4 Parental Competencies Parent’s creativity personality score 11.18 20 YEDU3A01w4 Relationship and School Life When I am struggling with studying or other issues, I would like to consult with the teacher first. 11.17 5 Discussion This study provides empirical evidence that adolescents’ socio-emotional competence is not fixed or developed in a linear trajectory but rather unfolds as a dynamic process shaped by the interplay among cognitive, emotional, and social components within broader ecological contexts. The three latent profiles identified through Latent Profile Analysis (LPA)—“Holistically Competent,” “Emotionally Sensitive but Socially Withdrawn,” and “Cognitively Engaged but Emotionally Detached”—highlight the heterogeneity in socio-emotional development. Notably, the observed discrepancy between emotional understanding (e.g., empathy) and behavioral enactment (e.g., sociability) challenges the prevailing notion that socio-emotional competence can be captured along a single, unified dimension. The identification of these distinct profiles underscores the methodological strength of the present study. For instance, adolescents in the “Cognitively Engaged but Emotionally Detached” group scored at average levels on self-report measures but demonstrated significantly lower performance on the Reading the Mind in the Eyes Test (RMET), an objective measure of cognitive empathy. Had this study relied solely on self-report instruments, such hidden deficits in emotional functioning may have remained undetected. The integration of both self-report and performance-based assessments was therefore essential to uncovering the complex—and at times contradictory—nature of socio-emotional competence. The results of the machine learning–based predictive analysis further corroborate Bronfenbrenner’s ecological model, revealing that variables within the adolescent’s microsystem exert the greatest influence on socio-emotional development. Among all ecological predictors, parental cognitive empathy emerged as the most dominant and consistent predictor, surpassing all other variables in predictive strength. This finding represents one of the study’s most novel and policy-relevant contributions, suggesting that parents' empathic dispositions form a foundational ecological bedrock for adolescents' socio-emotional growth. Theoretically, this study contributes by providing a practical and scalable approach to empirically operationalizing the OECD’s Study on Social and Emotional Skills (SSES) framework using nationally representative panel data. In doing so, it offers a methodological model for translating complex theoretical constructs into policy-relevant indicators, particularly within the constraints of applied research settings. This aligns closely with the core objectives of the field of child indicator research—namely, to inform child well-being policy through rigorous, multidimensional measurement. The findings yield several actionable implications for policy interventions aimed at enhancing adolescents’ socio-emotional skills. First, a shift in intervention focus from students to parents is warranted. The fact that parental cognitive empathy was over 3.8 times more predictive than any school- or student-level factor suggests that investing limited policy resources in parent-targeted empathy training and counseling programs may be the most cost-effective strategy. Second, a one-size-fits-all approach to intervention is unlikely to be effective. The discovery of three qualitatively distinct socio-emotional profiles points to divergent needs across subgroups. For example, adolescents in the “Emotionally Sensitive but Socially Withdrawn” profile may not benefit from generic empathy-building curricula, but instead require targeted social skills training to help them translate emotional sensitivity into behavioral expression. Despite its contributions, this study is not without limitations. First, the use of cross-sectional data restricts the ability to draw causal inferences between parental empathy and adolescents’ socio-emotional competencies. Second, due to structural constraints of large-scale panel data, certain subdomains (e.g., perseverance, sociability) were assessed using a limited number of items—a trade-off necessary to balance practical feasibility with international comparability. Future research should address these limitations by employing longitudinal designs to examine the developmental trajectories of socio-emotional profiles and test causal pathways involving ecological predictors. Additionally, expanding the scope to include macrosystem-level factors such as educational policy and cultural values will be essential for building a more comprehensive, multilayered model of socio-emotional development. 6 Conclusion Guided by the OECD Study on Social and Emotional Skills (SSES) framework, this study aimed to identify distinct socio-emotional competence (SEC) profiles among Korean adolescents and to uncover their key ecological predictors using machine learning techniques. The findings revealed that adolescent SEC does not follow a single, uniform trajectory; rather, it manifests across three qualitatively distinct subgroups—“Holistically Competent and Well-Adjusted,” “Emotionally Sensitive but Socially Withdrawn,” and “Cognitively Engaged but Emotionally Detached.” These results highlight the non-linear and heterogeneous nature of competence development during adolescence and underscore the need for differentiated policy approaches tailored to the specific strengths and vulnerabilities of each subgroup. Among the study’s most significant contributions is the empirical identification of parental cognitive empathy as the single most powerful and dominant ecological predictor of adolescents’ socio-emotional competence. Among dozens of individual-, family-, and school-level variables, this factor surpassed all others in predictive strength. This finding has critical implications for public policy, suggesting that fostering empathic capacities in parents may serve as the most effective leverage point for promoting healthy adolescent development. In conclusion, this study offers both theoretical and methodological contributions to the field of child indicator research, while simultaneously providing robust empirical evidence to inform practice. By demonstrating that parental empathy is a central ecological driver of adolescent socio-emotional outcomes, the findings establish a clear and actionable direction for designing impactful and resource-efficient interventions. Declarations Author Contribution The sole author was responsible for the conception and design of the study, data analysis and interpretation, as well as the drafting and revision of the manuscript. References Blakemore, S., & Mills, K. (2014). Is adolescence a sensitive period for sociocultural processing?. Annual Review of Psychology, 65 , 187-207. Breiman, L. (2001). Random forests. Machine learning, 45 (1), 5-32. Bronfenbrenner, U. (1994). Ecological models of human development . In International Encyclopedia of Education. Oxford: Elsevier. Brown, L., Green, K., & Taylor, D. (2021). Coping with COVID-19: Adolescent stress and coping strategies. Journal of Youth and Adolescence, 50 (5), 987-1001. Burke, J. D., & Loeber, R. (2016). Mechanisms of behavioral and affective treatment outcomes in a cognitive behavioral intervention for boys. Journal of Abnormal Child Psychology, 44 (1), 179–189. Casey, B. (2015). Beyond simple models of self-control to circuit-based accounts of adolescent behavior. Annual Review of Psychology, 66 , 295-319. Chernyshenko, O., M. Kankaraš and F. Drasgow (2018). Social and emotional skills for student success and well-being: Conceptual framework for the OECD study on social and emotional skills. OECD Education Working Papers, 173, OECD Publishing: Paris. Collaborative for academic, social, and emotional learning. (2015). 2015 CASEL guide: Effective social and emotional learning programs – Middle and high school edition. Chicago, IL: Author. Datu, J., & Restubog, S. (2020). The emotional pay-off of staying gritty: linking grit with social-emotional learning and emotional well-being. British Journal of Guidance & Counselling, 48 (5), 697–708. Durlak, J., Weissberg, R., Dymnicki, A., Taylor, R., & Schellinger, K. (2011). The impact of enhancing students’ social and emotional learning: A meta-analysis of school-based universal interventions. Child Development, 82 (1), 405-432. Glick, G., & Rose, A. (2011): Prospective associations between friendship adjustment and social strategies: Friendship as a context for building social skills. Developmental Psychology, 47 (4), 1117-1132. Ha, H., Choi, Y., Jeong, E., Jeong, Y., & Han, J. (2017). Korean Children and Youth Panel Survey VIII: Project report. National Youth Policy Institute. Han, S., Kim, H., Seol, I., Lim, Y., & Cho, A. (2014). Adolescent Psychology (2nd ed.). Gyeonggi: Education Science Press. Huhtala, M., Korja, R., Lehtonen, L., Haataja, L., Lapinleimu, H., Rautava, P., & PIPARI Study Group. (2014). Associations between parental psychological well-being and socio-emotional development in 5-year-old preterm children. Early Human Development, 90 (3), 119-124. Jedidi, K., Jagpal, H. S., & DeSarbo, W. (1997). Finite-mixture structural equation models for response-based segmentation and unobserved heterogeneity. Marketing Science, 16 (1), 39-59. Jones, G., White, R., & Smith, P. (2022). Effects of remote learning on social-emotional development in adolescents. Educational Psychology Review, 34 (1), 123-145. Jung, T., & Wickrama, K. A. (2008). An introduction to latent class growth analysis and growth mixture modeling. Social and personality psychology compass, 2 (1), 302-317. Kim, H., Kim, M., Lee, S., Yang, H., Kim, J., Kim, M., Kim, J., & Cho, Y. (2020). OECD ESP Social and Emotional Skills Survey International Collaborative Study (IV): Analysis of Main Survey Results. Chungbuk: Korea Educational Development Institute. Kim, K., Kim, W., Choi, I., Kim, M., Kim, K., Park, J., Kim, S., Park, J., Do, S., & Jang, H. (2020). Development of student competency assessment tool for the 2020 Seoul education longitudinal study. Korea Institute for Curriculum and Evaluation. Research Report. Martinez-Yarza, N., Santibáñez, R., & Solabarrieta, J. (2023). A Systematic Review of Instruments Measuring Social and Emotional Skills in School-Aged Children and Adolescents. Child Indicators Research, 16 (4), 1475–1502. Mishra, N., & Singh, P. (2021). Feature construction and smote-based imbalance handling for multi-label learning. Information Sciences, 563 , 342-357. Napolitano, C. M., Greenberg, M. T., & Weissberg, R. P. (2021). Toward a Developmental Framework of Social-Emotional Competence for Youth. American Psychologist, 76 (6), 878–893. OECD (2015). Skills for social progress: The power of social and emotional skills. OECD Publishing. Orth, U., Robins, R., & Widaman, K. (2012). Life-span development of self-esteem and its effects on important life outcomes. Journal of Personality and Social Psychology, 102 (6), 1271-1288. Peetz, H. K., Lansu, T. A., Hoekstra, N. A., van den Berg, Y. H., Burk, W. J., & Mainhard, M. T. (2025). Assessing youth's internal and external attributions to negative peer interactions and victimization—development of the Causal Attributions for Peer Experiences (CAPE) scale. Journal of Research on Adolescence , 35 (2), e70037. Portela-Pino, I., Alvariñas-Villaverde, M., & Pino-Juste, M. (2021). Socio-emotional skills in adolescence. Influence of personal and extracurricular variables. International Journal of Environmental Research and Public Health, 18 (9), 4811. Roorda, D., Koomen, H., Spilt, J., & Oort, F. (2011). The influence of affective teacher–student relationships on students’ school engagement and achievement: A meta-analytic approach. Review of Educational Research, 81 (4), 493-529. Ross, K. M., Kim, H., Tolan, P. H., & Jennings, P. A. (2019). An exploration of normative social and emotional skill growth trajectories during adolescence. Journal of Applied Developmental Psychology, 62 , 151–162. Salmela-Aro, K., & Upadyaya, K. (2020). School engagement and school burnout profiles during high school–The role of socio-emotional skills. European Journal of Developmental Psychology, 17 (6), 943-964. Savickas, M. (2013). Career construction theory and practice. Career development and counseling. Putting theory and research to work, 2, 144-180. Sklad, M., Diekstra, R., Ritter, M. D., Ben, J., & Gravesteijn, C. (2012). Effectiveness of school‐based universal social, emotional, and behavioral programs: Do they enhance students’ development in the area of skill, behavior, and adjustment?. Psychology in the Schools, 49 (9), 892-909. Smith, A., Brown, C., & Jones, E. (2021). The impact of the COVID-19 pandemic on adolescent mental health. Journal of Adolescent Health, 68 (3), 456-463. Stekhoven, D. J., & Bühlmann, P. (2012). MissForest—non-parametric missing value imputation for mixed-type data. Bioinformatics, 28 (1), 112-118. Thompson, R. B., Corsello, M., McReynolds, S., & Conklin-Powers, B. (2013). A Longitudinal study of family socioeconomic status (SES) variables as predictors of socio-emotional resilience among mentored youth. Mentoring & Tutoring: Partnership in Learning, 21 (4), 378–391. Yu, M., & Lee, K. (2023). Study on the implementation of the 2023 Convention on the Rights of the Child: Basic analysis report on the status of children’s and adolescents’ rights in Korea. National Youth Policy Institute. Woo, Y., Lee, J., Lee, I., Hong, W., Choi, J., Jung, E., & Kwon, Y. (2022). Development of diagnostic tools and guidance/support materials for socio-emotional competence. Korea Institute for Curriculum and Evaluation, Research Report. Additional Declarations No competing interests reported. Supplementary Files Appendix.docx Cite Share Download PDF Status: Published Journal Publication published 13 Mar, 2026 Read the published version in Child Indicators Research → 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-7891642","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":540663685,"identity":"f06cd313-27ac-4392-8243-0e2093bcb8a9","order_by":0,"name":"Nayoung Kim","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2klEQVRIiWNgGAWjYJACCR4DIMneACQMLEjRwnMApEWCWC1gMgFMElbOP7v94o03BXZ58pHPr274USDBwN/enYDfhjtnii3nGCQXG97OKbvZA3SYxJmzG/BbcyMnTZrHgDlx4+yctBs8QC0GErn4tchDtNQnbpx5Ju3mH2K0GNxIPwbUcjhxvgT7sdtE2WJ4I4cZ6JfjiRt4cthuyxhI8BD0i9yN9Ic33vypTpzffvzZzTd/bOT423sJeJ8BHI9AFx6AMHgIKAcB9gdgSr4ByhgFo2AUjIJRgA4AcjVKBhOLH/sAAAAASUVORK5CYII=","orcid":"","institution":"Korean Educational Development Institute","correspondingAuthor":true,"prefix":"","firstName":"Nayoung","middleName":"","lastName":"Kim","suffix":""}],"badges":[],"createdAt":"2025-10-18 07:23:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7891642/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7891642/v1","draftVersion":[],"editorialEvents":[{"content":"https://doi.org/10.1007/s12187-026-10355-w","type":"published","date":"2026-03-13T15:58:27+00:00"}],"editorialNote":"","failedWorkflow":false,"files":[{"id":95807506,"identity":"535e3ebf-cc34-422d-bd82-93539dca1b33","added_by":"auto","created_at":"2025-11-13 08:48:50","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":400198,"visible":true,"origin":"","legend":"","description":"","filename":"BlindedManuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/184f2c01a20d270b3bfb483b.docx"},{"id":95807695,"identity":"6a2daa20-e946-4169-9d91-9ed123b4dad5","added_by":"auto","created_at":"2025-11-13 08:49:05","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3477,"visible":true,"origin":"","legend":"","description":"","filename":"6783b11c80444d51a2f988e02fc578bf.json","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/527e6ffe3fe4e0e40fddbb49.json"},{"id":95807670,"identity":"5d77253b-0293-4ad9-8a80-0821b700bab4","added_by":"auto","created_at":"2025-11-13 08:49:03","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":124608,"visible":true,"origin":"","legend":"","description":"","filename":"6783b11c80444d51a2f988e02fc578bf1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/29c6e5f8d3aad4e0c767cc41.xml"},{"id":95808091,"identity":"c143d8bf-8b39-47d0-91de-fc95d08bab15","added_by":"auto","created_at":"2025-11-13 08:49:19","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48431,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/71d6a0e5534949672670491e.png"},{"id":95807709,"identity":"2176b729-7827-4dc2-bb90-b1d5e1529d11","added_by":"auto","created_at":"2025-11-13 08:49:06","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":288577,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/15a548d7bdcc0d5454b7a585.png"},{"id":95808076,"identity":"4f7c617d-a237-4e9d-82e4-09e154b09651","added_by":"auto","created_at":"2025-11-13 08:49:18","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":14986,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/a76395a6dfcf8e1dd9ccdf0e.png"},{"id":95808090,"identity":"aa274a20-1389-4276-8e52-09575dd09928","added_by":"auto","created_at":"2025-11-13 08:49:19","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":77087,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/c2067aea71da895f679d9ad8.png"},{"id":95807573,"identity":"d1ce22eb-6eed-4a9c-aa9e-c5719379f161","added_by":"auto","created_at":"2025-11-13 08:48:57","extension":"xml","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":122500,"visible":true,"origin":"","legend":"","description":"","filename":"6783b11c80444d51a2f988e02fc578bf1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/74a789f528bf877fc9b3074f.xml"},{"id":95808005,"identity":"568d30c5-00b2-4504-8b12-df0a46df2f51","added_by":"auto","created_at":"2025-11-13 08:49:16","extension":"html","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":128599,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/afd15bbefa42060dc94ede17.html"},{"id":95807951,"identity":"41f1b2cb-8423-4488-8c29-8fe3b7716251","added_by":"auto","created_at":"2025-11-13 08:49:16","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":103540,"visible":true,"origin":"","legend":"\u003cp\u003eStandardized scores of five socio-emotional indicators by latent profile\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNote.\u003c/strong\u003e While empathy showed the largest difference between profiles, meaningful variation was also observed in sociability (range: –0.4 to +0.63), perseverance (–0.43 to +0.73), and creativity (–0.11 to +0.78), indicating that classification was multidimensional.\u003c/p\u003e","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/87fc519d7c9033dd02e7c822.png"},{"id":104739748,"identity":"0f58ffe4-738c-41ec-b475-d5e858014c6f","added_by":"auto","created_at":"2026-03-16 16:12:37","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1025060,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/d81bae99-c070-4131-bb2f-3b8cb9d54024.pdf"},{"id":95807593,"identity":"219e3861-c106-4e04-9c82-934d3457d14e","added_by":"auto","created_at":"2025-11-13 08:48:59","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":304236,"visible":true,"origin":"","legend":"","description":"","filename":"Appendix.docx","url":"https://assets-eu.researchsquare.com/files/rs-7891642/v1/e03f5dbfe06c5b6cc2179c09.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Ecological and Predictive Indicators of Social and Emotional Skills among Korean Adolescents: A Person-Centered Analysis within the OECD SSES Framework","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eAdolescents today are growing up amid rapid transformations in social structures and value systems. The aftermath of the COVID-19 pandemic, marked by prolonged social isolation and emotional instability, has posed serious challenges to their development and highlighted the urgent need for robust data to guide policy responses (Brown et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Jones et al., \u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e2022\u003c/span\u003e; Smith et al., \u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Within this context, socio-emotional skills (SoES)—the abilities to regulate emotions, empathize with others, and make responsible decisions—have emerged not only as critical developmental assets but also as a crucial indicator for monitoring national child well-being (Blakemore \u0026amp; Mills, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Casey, \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). As foundational competencies for resilience and social adaptation, these indicators are essential for policymakers seeking to design effective support systems and evaluate the psychological health of the youth population. Understanding the structure and predictors of SoES is therefore a prerequisite for evidence-based adolescent policy and practice (Durlak et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2011\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo conceptualize and promote these skills, various frameworks have been proposed. Among them, the CASEL (Collaborative for Academic, Social, and Emotional Learning) model has become a dominant paradigm, offering a systematic structure for implementing social and emotional learning (SEL) in schools. It comprises five core competencies—self-awareness, self-management, social awareness, relationship skills, and responsible decision-making (CASEL, 2015)—and has gained widespread adoption due to its strong curricular applicability. However, despite its practical utility, the CASEL model has been critiqued for its relatively static and linear conceptualization of learners’ internal competencies, which may overlook the contextual and dynamic nature of socio-emotional development.\u003c/p\u003e\u003cp\u003eEmpirical evidence supports this critique. For example, Ross et al. (\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e) demonstrated through a large-scale longitudinal study that adolescents’ socio-emotional skills develop along diverse and non-linear trajectories that vary by skill domain and gender. Some adolescents may exhibit strong empathy yet struggle to maintain peer relationships, whereas others may be socially outgoing but less adept at emotional regulation. Such asynchronous developmental patterns cannot be fully captured through variable-centered approaches, which emphasize mean-level associations. Accordingly, scholars have highlighted the importance of person-centered approaches that uncover heterogeneity and within-person configurations of multiple competencies (Blakemore \u0026amp; Mills, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Ross et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2019\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe person-centered perspective offers both theoretical and practical advantages by focusing on how different competencies interact within individuals rather than examining them in isolation. Latent Profile Analysis (LPA), in particular, enables researchers to identify statistically distinct subgroups based on patterns of socio-emotional functioning and to examine how these profiles relate to developmental outcomes (Jung \u0026amp; Wickrama, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). For instance, Ağırkan and Haspolat (2025) applied LPA to classify high-school students’ socio-emotional competence profiles and reported meaningful differences in academic adjustment and psychological well-being across subgroups, thereby underscoring the multidimensional and differentiated nature of socio-emotional competence.\u003c/p\u003e\u003cp\u003eBuilding on this person-centered foundation, the present study adopts the OECD Social and Emotional Skills (SSES) framework to explore the heterogeneity of SoES within a developmental-ecological context. In contrast to practitioner-focused models like CASEL, the OECD SSES provides a robust theoretical model for understanding socio-emotional development as a dynamic, contextually embedded process (Chernyshenko et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; OECD, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). This ecological emphasis is rooted in its integration of the Big Five personality theory with Bronfenbrenner’s ecological systems theory (1994), which posits that competencies are malleable resources shaped by continuous interactions across individual, family, and school systems (Lerner et al., 2018). A comprehensive understanding based on this framework, therefore, inherently demands analytical approaches that can capture both intra-individual heterogeneity (i.e., how skills combine within a person) and inter-contextual dynamics (i.e., how environmental factors predict these combinations). Guided by this perspective, the study operationalizes the SSES framework—which has demonstrated strong cross-cultural validity (Kim, H., et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Kim, K., et al., \u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Woo et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2022\u003c/span\u003e)—to classify adolescents’ socio-emotional profiles and identify key ecological predictors.\u003c/p\u003e\u003cp\u003ePrevious Korean research has largely relied on variable-centered methods, offering limited insight into how multiple competencies interact within ecological contexts. To address this gap, the present study employs a hybrid design integrating Latent Profile Analysis (LPA) with a stacked ensemble machine learning model (CAWPE). This combination enables both robust classification of socio-emotional profiles and identification of influential ecological predictors. By merging developmental-ecological theory with predictive analytics, this study advances a more comprehensive understanding of the dynamics underlying adolescent SoES. Ultimately, this research provides a methodological foundation to move beyond one-size-fits-all interventions, enabling the design of evidence-based, targeted policies aimed at the specific needs of distinct adolescent subgroups. The study addresses the following research questions:\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003col\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eWhat latent profiles characterize adolescents’ socio-emotional skills?\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003cspan\u003e\u003cli\u003e\u003cp\u003eWhich ecological factors at the individual, family, and school levels significantly predict membership in each profile?\u003c/p\u003e\u003c/li\u003e\u003c/span\u003e\u003c/ol\u003e\u003cp\u003e\u003c/p\u003e"},{"header":"2 Literature Review","content":"\u003ch2\u003e2.1 Conceptualizing Socio-Emotional Skills\u003c/h2\u003e\u003cp\u003eSocio-emotional skills (SoES) represent essential developmental assets that foster adolescents’ psychological well-being, interpersonal competence, and academic adjustment. These skills encompass a broad spectrum of social and emotional abilities—including emotional regulation, empathy, collaboration, and responsible decision-making—that enable adaptive functioning within social contexts (Blakemore \u0026amp; Mills, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Durlak et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Rather than being innate traits, SoES are best understood as developmental capacities that emerge through ongoing interactions between individual dispositions (e.g., temperament, personality) and environmental influences (OECD, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). Thus, socio-emotional skills are dynamic, context-sensitive competencies that evolve across time and experience.\u003c/p\u003e\u003cp\u003eThe OECD Social and Emotional Skills Framework (SSES) offers a comprehensive theoretical model for conceptualizing this developmental process. Grounded in the Big Five personality structure—openness, conscientiousness, extraversion, agreeableness, and emotional stability—the SSES reconceptualizes personality traits as malleable socio-emotional competencies (Chernyshenko et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e). It categorizes these competencies into five broad domains: task performance, emotional regulation, collaboration, open-mindedness, and engagement with others. Each domain comprises specific subskills such as perseverance, optimism, empathy, and sociability, which have been empirically linked to adolescents’ academic success, psychological resilience, and social adjustment.\u003c/p\u003e\u003cp\u003eA distinctive strength of the SSES lies in its dual perspective: it views socio-emotional skills as both personality-based psychological resources and plastic, context-responsive competencies. This approach contrasts with the widely adopted CASEL (Collaborative for Academic, Social, and Emotional Learning) model, which defines five core competencies—self-awareness, self-management, social awareness, relationship skills, and responsible decision-making (CASEL, 2015). While the CASEL model provides a practical framework for implementing school-based SEL programs, it tends to conceptualize these competencies as static functional domains. Recent systematic reviews of SEL measurement instruments (Martinez-Yarza, et al., \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) have highlighted that the majority of existing tools are grounded in the CASEL framework, with limited attention to alternative conceptualizations such as the OECD SSES. This imbalance underscores the need for frameworks that better capture the multidimensional and context-sensitive nature of socio-emotional skills. The SSES, in contrast, expands this understanding by embedding socio-emotional skills within a developmental–ecological perspective, emphasizing within-person heterogeneity and cross-contextual interdependence (Lerner et al., 2018; OECD, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). To address the limitations inherent in the traditional variable-centered approach, recent scholarship has increasingly adopted a person-centered perspective in examining adolescents’ socio-emotional skills (SoES). In particular, research employing Latent Profile Analysis (LPA) has empirically demonstrated that adolescent SoES are not monolithic, but instead are expressed through qualitatively distinct and heterogeneous subgroups (e.g., Peetz et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2025\u003c/span\u003e). Nevertheless, current LPA-based studies exhibit several notable constraints: (a) they are primarily grounded in the CASEL theoretical framework; (b) they focus predominantly on Western populations, limiting cultural generalizability; and (c) they have seldom utilized integrative frameworks like the OECD’s Study on Social and Emotional Skills (SSES) to explore the multidimensional nature of socio-emotional profiles in conjunction with their broader ecological contexts.\u003c/p\u003e\u003cp\u003eThe SSES framework aligns closely with Bronfenbrenner’s (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) ecological systems theory, which posits that development arises from reciprocal interactions between individuals and nested environmental systems—such as family, peers, schools, and communities. Extending this ecological foundation, the SSES conceptualizes socio-emotional competencies as multi-layered and adaptive resources shaped by continuous exchanges across ecological levels. In this regard, the SSES functions not merely as a personality model but as an applied developmental framework that integrates insights from developmental psychology and ecology to explain the dynamic and interactive nature of socio-emotional growth.\u003c/p\u003e\u003cp\u003eIn Korea, the increasing policy emphasis on social and emotional learning (SEL) has led to the integration of SEL components into the national curriculum and to growing interest in developing culturally grounded measurement instruments. Recent studies by Kim K., et al. (\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e2020\u003c/span\u003e) and Woo et al. (\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) have validated the structural and cultural validity of the OECD SSES framework, confirming its applicability for capturing the socio-emotional characteristics and developmental diversity of Korean adolescents. These findings demonstrate that the SSES serves not only as a measurement tool but also as a conceptual and analytical lens for examining socio-emotional development across cultural contexts.\u003c/p\u003e\u003cp\u003eBuilding upon this theoretical and empirical foundation, the present study extends the application of the SSES framework to a non-Western OECD context, empirically operationalizing its core indicators within nationally representative data. By integrating person-centered analysis (Latent Profile Analysis) with ecological and predictive modeling, this study advances the measurement and understanding of adolescent socio-emotional skills as multidimensional, dynamic, and contextually embedded phenomena—thereby addressing a key gap in comparative child indicator research.\u003c/p\u003e\u003ch2\u003e2.2 Multilevel Predictors of Socio-Emotional Skills\u003c/h2\u003e\u003cp\u003eAdolescents’ socio-emotional skills (SoES) are shaped not by a single factor but through a multilevel ecological developmental process in which individual, familial, school, and community systems interact dynamically (Bronfenbrenner, \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1994\u003c/span\u003e). Within this framework, the development of SoES is driven primarily by proximal processes—the everyday relational interactions with parents, teachers, and peers—rather than by distal structural or institutional influences. Integrating the OECD Social and Emotional Skills (SSES) framework with Bronfenbrenner’s ecological model, this study conceptualizes SoES as ecologically embedded and measurable developmental indicators, rather than as fixed psychological traits.\u003c/p\u003e\u003cp\u003e\u003cb\u003eIndividual-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAt the individual level, psychological resources such as self-esteem, emotional stability, aggression, and impulsivity serve as foundational predictors of socio-emotional competence (Burke \u0026amp; Loeber, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2016\u003c/span\u003e; Orth et al., \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). High self-esteem and emotional stability foster empathy, prosocial behavior, and resilience, whereas aggression and impulsivity undermine emotional regulation and interpersonal adjustment. In addition, lifestyle and self-regulatory behaviors—including adequate sleep, regular physical activity, and balanced media use—operate as moderators that promote emotional well-being and facilitate positive peer relationships (Portela-Pino et al., \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2021\u003c/span\u003e; Yu \u0026amp; Lee, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). Academic variables also play a key role: academic achievement and engagement strengthen motivation and self-regulation, while academic stress and burnout have been linked to decreased socio-emotional functioning (Salmela-Aro \u0026amp; Upadyaya, \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2020\u003c/span\u003e). Collectively, these individual-level indicators align with the task performance and emotional regulation domains of the SSES framework.\u003c/p\u003e\u003cp\u003e\u003cb\u003eFamily-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe family constitutes the primary context for emotional socialization during adolescence. Parental emotional support—characterized by warmth, empathy, and responsiveness—directly enhances adolescents’ emotional regulation and interpersonal trust (Datu \u0026amp; Restubog, \u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Ong et al., 2018). Beyond relational dynamics, structural family resources such as parental education and household income provide cognitive and affective stimulation that fosters socio-emotional growth (Thompson et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2013\u003c/span\u003e). Parental cognitive empathy and creativity, in particular, contribute to adolescents’ open-mindedness and problem-solving capacities by modeling reflective and flexible thinking. Moreover, career-related parent–child interactions, including shared goal setting and discussions of future plans, strengthen autonomy and future orientation—skills conceptually situated within the task performance and open-mindedness domains of the SSES (Huhtala et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2014\u003c/span\u003e). Thus, family-level factors jointly represent both structural resources and interactive mechanisms that cultivate adolescents’ socio-emotional competencies.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSchool-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eThe school serves as a core microsystem for socialization, where adolescents internalize social norms and interactional skills through daily engagement with teachers and peers. The quality of teacher–student relationships is consistently associated with school belonging, learning motivation, and emotional regulation (Roorda et al., \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Similarly, peer acceptance and supportive friendships act as protective buffers that mitigate externalizing behaviors and social withdrawal (Glick \u0026amp; Rose, \u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2011\u003c/span\u003e). Participation in extracurricular and self-directed activities—such as clubs, mentoring programs, and career exploration—further reinforces cooperation, empathy, and self-control, contributing to both academic achievement and long-term social responsibility (Durlak et al., \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e2011\u003c/span\u003e; Han et al., \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e2014\u003c/span\u003e; Sklad et al., \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2012\u003c/span\u003e). Furthermore, experiences of career counseling and identity exploration have been identified as salient predictors of socio-emotional growth, promoting emotional maturity and goal-directed self-regulation (Savickas, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e2013\u003c/span\u003e; Napolitano et al., \u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Collectively, these school-level indicators correspond to the collaboration and engagement with others domains of the SSES, underscoring the school’s role as an everyday ecological context for the development of emotional and interpersonal competencies.\u003c/p\u003e\u003cp\u003eIn conclusion, existing research offers compelling evidence that factors at the individual, familial, and school levels are significant ecological predictors of students' sense of emotional and social well-being (SoES). Nevertheless, much of this work has predominantly employed traditional regression methods that examine the linear and independent contributions of each predictor. Such analytical approaches are often inadequate for capturing the intricate, non-linear, and interactive dynamics characteristic of complex ecological systems. As a result, they fall short in addressing a fundamental policy-relevant question: which among these numerous factors exert the greatest influence on adolescent developmental outcomes? This methodological limitation underscores the need for more sophisticated analytical techniques. The current study responds to this gap by adopting a predictive analytics framework, specifically machine learning, to not only corroborate established ecological associations but also to identify the most influential indicators of subgroup membership. This approach advances beyond mere correlational analysis, providing a more robust and actionable evidence base to inform targeted policy interventions.\u003c/p\u003e"},{"header":"3 Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Data and Participants\u003c/h2\u003e\u003cp\u003eThis study drew on data from the fourth wave (2021) of the Korean Children and Youth Panel Survey (KCYPS) 2018, conducted by the National Youth Policy Institute (NYPI) of Korea. The KCYPS is a nationally representative longitudinal study designed to examine developmental, psychosocial, and educational indicators across key life transitions in childhood and adolescence (Ha et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e). The survey includes two cohorts\u0026mdash;one initiated in the fourth grade of elementary school and another in the first grade of middle school\u0026mdash;and provides rich information on adolescents\u0026rsquo; socio-emotional, behavioral, and contextual characteristics.\u003c/p\u003e\u003cp\u003eThe present study focused on the elementary school cohort, who were in their first year of middle school (Grade 7) at the time of the fourth wave. This developmental stage represents early adolescence, a critical transition period during which socio-emotional skills and self-regulatory competencies undergo rapid change. While the KCYPS has a longitudinal design, the current analysis used cross-sectional data from Wave 4, as the socio-emotional skills (SoES) indicators were uniquely collected in a special module during this wave. Consequently, longitudinal analyses (e.g., intra-individual growth trajectories) were not feasible. Nonetheless, the present design offers a valuable baseline validation of socio-emotional skill indicators within a nationally representative dataset, serving as an empirical foundation for future longitudinal research.\u003c/p\u003e\u003cp\u003eThe initial sample comprised 2,607 adolescents. After excluding participants with missing values on core SoES indicators, the final analytic sample consisted of 2,275 students (50.3% male, n\u0026thinsp;=\u0026thinsp;1,145; 49.7% female, n\u0026thinsp;=\u0026thinsp;1,130). Attrition analysis indicated no significant group differences in demographic and ecological variables between retained and excluded participants, suggesting that the final sample remained demographically representative of the original cohort.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Measures\u003c/h2\u003e\u003cp\u003e\u003cb\u003eSocial and Emotional Skills\u003c/b\u003e\u003c/p\u003e\u003cp\u003eAdolescents\u0026rsquo; socio-emotional skills (SoES) were defined and measured in accordance with the OECD Social and Emotional Skills Framework (SSES). The SSES conceptualizes socio-emotional skills as psychological resources that develop through continuous interaction between personality-based dispositions and environmental factors, encompassing five core domains: (1) task performance, (2) emotional regulation, (3) collaboration, (4) open-mindedness, and (5) engagement with others (Chernyshenko et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; OECD, \u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e2015\u003c/span\u003e). While the OECD SSES covers a broad and multidimensional construct space, it is \u003cem\u003emethodologically impractical to include all subdomains within a single survey\u003c/em\u003e. Therefore, this study operationalized the SSES framework by selecting theoretically representative and empirically validated indicators available in the Korean Children and Youth Panel Survey (KCYPS). Each selected variable was mapped to the SSES domains based on prior validation studies (Kim, H., et al., \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e2020\u003c/span\u003e; Woo et al., \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e2022\u003c/span\u003e) and its conceptual correspondence to the five core areas of socio-emotional competence. This process ensured that the operationalization achieved a representative balance across domains, maintaining the structural integrity and conceptual breadth of the SSES rather than focusing narrowly on any single subfactor.\u003c/p\u003e\u003cp\u003eSpecifically, indicators included perseverance (4 items, Cronbach's α\u0026thinsp;=\u0026thinsp;.610) for task performance, optimism (5 items, α\u0026thinsp;=\u0026thinsp;.822) for emotional regulation, cognitive empathy (as assessed by the Reading the Mind in the Eyes Test) for collaboration (28 items, α\u0026thinsp;=\u0026thinsp;.724), creative personality traits (30 items, α\u0026thinsp;=\u0026thinsp;.707) for open-mindedness, and sociability (5 items, α\u0026thinsp;=\u0026thinsp;.887) for engagement with others. All items were standardized (z-scores) prior to analysis to enable direct comparison across measurement scales and domains. The subdomains, representative indicators, and reliability coefficients (Cronbach\u0026rsquo;s α) used in this study are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\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\u003eMeasures of socio-emotional skills (SSES Framework)\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=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eDomain\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRepresentative Subskills\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eExample Item\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCronbach\u0026rsquo;s α (Number of items)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eTask Performance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePerseverance\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ldquo;I don\u0026rsquo;t give up easily even when things are difficult.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.610 (4)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEmotional Regulation\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eOptimism\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ldquo;I tend to think positively about the future.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.822(5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eCollaboration\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCognitive empathy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eScore converted from responses to 28 items in the Reading the Mind in the Eyes Test (RMET), representing emotional recognition ability.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.724(28)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eOpen-Mindedness\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCreativity\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ldquo;I tend to be creative in my thinking and problem-solving.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.707(30)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eEngagement with Others\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSociability\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e\u0026ldquo;I find it easy to get along with others.\u0026rdquo;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.887(5)\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\u003cb\u003ePredictor Variables\u003c/b\u003e\u003c/p\u003e\u003cp\u003eTo identify the multilevel ecological predictors of adolescents\u0026rsquo; socio-emotional competence, this study organized the independent variables within a theoretically grounded framework. The selection of predictors was informed by the OECD Social and Emotional Skills (SSES) framework, which integrates the Big Five personality model with Bronfenbrenner\u0026rsquo;s (\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e1994\u003c/span\u003e) ecological systems theory... (Chernyshenko et al., \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2018\u003c/span\u003e; Lerner et al., 2018). The inclusion of a comprehensive set of variables spanning these ecological domains was an intentional methodological strategy. This approach was designed to leverage the primary strength of machine learning: its ability to sift through a high-dimensional and complex feature space and identify the most salient predictive indicators in a data-driven manner, moving beyond the limitations of traditional regression models. Consistent with this dual theoretical and analytical strategy, the predictors were systematically categorized across three ecological levels: individual, family, and school. The specific variables included within each of these levels are described in detail below.\u003c/p\u003e\u003cp\u003e\u003cem\u003eIndividual-Level Factors\u003c/em\u003e\u003c/p\u003e\u003cp\u003eAt the individual level, predictors were organized into four categories. The first category, psychological attributes, included self-esteem, emotional stability, aggression, attention, happiness, and depressive symptoms. The second category, behavioral and lifestyle variables, comprised average sleep duration, physical activity frequency, smartphone usage, and leisure time. The third category, academic and career-related factors, encompassed academic achievement, academic engagement, academic helplessness, career adaptability, and family discussions about career planning. Finally, delinquent behavior indicators, such as experiences of offline or online delinquency, were included.\u003c/p\u003e\u003cp\u003e\u003cem\u003eFamily-Level Factors\u003c/em\u003e\u003c/p\u003e\u003cp\u003eAt the family level, predictors were grouped into three categories. The first, socio-demographic characteristics, included parental education, household income, and weekly working hours. The second, parental psychological resources, encompassed parents' cognitive empathy and creativity scores. The third category, parenting and relational dynamics, consisted of the frequency of parent-child communication, parental satisfaction with the child's academic progress, and parents' future educational expectations for their child.\u003c/p\u003e\u003cp\u003e\u003cem\u003eSchool-Level Factors\u003c/em\u003e\u003c/p\u003e\u003cp\u003eAt the At the school level, predictors were divided into two categories. The first, structural conditions, included school location (urban/rural) and the average number of instructional days per week. The second category, relational and psychological factors, comprised peer relationship quality, teacher-student relationship quality, and school satisfaction. A complete list of all predictive variables used in the machine learning model, along with their operational definitions and measurement scales, is summarized in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\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\u003ePredictive variables used in the classification of latent profiles\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\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\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c2\" namest=\"c1\"\u003e\u003cp\u003eDomain\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePredictive Variables (Number of items)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"4\" rowspan=\"5\"\u003e\u003cp\u003eStudents\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eGender, sleep duration, leisure time, duration of physical activity, smartphone usage, smartphone dependency, annual participation frequency in various activities\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePhysical and Psychological Characteristics\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePhysical symptoms, health status, attention, aggression, happiness, depression, self-esteem, and cooperativeness\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eLearning\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAcademic achievement, academic engagement, academic helplessness, study time, and participation in private tutoring\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCareer\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eCareer adaptability, frequency of career-related conversations, and decision-making status regarding future occupation\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eDelinquent Behavior\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePresence and frequency of real-world delinquent experiences, cyber-delinquent experiences\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"3\" rowspan=\"4\"\u003e\u003cp\u003eHome\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eHome\u003c/p\u003e\u003cp\u003eBackground\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParents\u0026rsquo; highest level of education, monthly household income, economic status of the family, parents\u0026rsquo; weekly working hours\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eParental Competencies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eparents\u0026rsquo; cognitive empathy score, and parental creativity index\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eParent-Child Relationship\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParental attitudes, time spent with parents, and time spent in conversations with parents\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eParents\u0026rsquo; Educational Planning and Satisfaction with School\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eChildren\u0026rsquo;s satisfaction with school life, desired level of education for the future\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eSchool\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSchool Characteristics\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRegion, average number of school days per week\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePeer relationships, teacher relationships, and satisfaction with school life\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003ctfoot\u003e\u003ctr\u003e\u003ctd colspan=\"3\"\u003e\u003cb\u003eNote.\u003c/b\u003e Most continuous variables were measured on a 5-point Likert scale, while items related to competencies were measured on a 7-point scale\u003c/td\u003e\u003c/tr\u003e\u003c/tfoot\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e3.3 Analytical Strategy\u003c/h2\u003e\u003cp\u003eThis study adopted a person-centered, two-stage analytical design to identify heterogeneous profiles of adolescents\u0026rsquo; socio-emotional competencies and to determine their multilevel ecological predictors. The analytic framework combined latent profile analysis (LPA) for classification and ensemble-based machine learning for predictive modeling, thereby linking theory-driven typology with data-driven inference. In the first stage, LPA was employed to classify participants into latent subgroups based on five standardized socio-emotional indicators: creativity, perseverance, optimism, sociability, and empathy. LPA is a probabilistic clustering technique that identifies unobserved latent classes within a population and is particularly suited to detecting individual-level heterogeneity in psychological constructs. Model selection was guided by multiple fit indices, including the Akaike Information Criterion (AIC), Bayesian Information Criterion (BIC), sample-size adjusted BIC (SA-BIC), entropy, and the Lo\u0026ndash;Mendell\u0026ndash;Rubin likelihood ratio test (LMR-LRT).\u003c/p\u003e\u003cp\u003eA three-class solution was selected based on both statistical fit and theoretical interpretability, reflecting (1) Holistically Competent and Well-Adjusted, (2) Emotionally Sensitive but Socially Withdrawn, and (3) Cognitively Engaged but Emotionally Detached subgroups (see Tables\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e and \u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e). In the second stage, a supervised machine learning approach was used to examine the relative influence of individual, family, and school-level predictors on latent profile membership. Specifically, the Cross-validation Accuracy Weighted Probabilistic Ensemble (CAWPE) algorithm was applied (Large et al., 2019). CAWPE integrates predictions from multiple base classifiers\u0026mdash;Random Forest, Gradient Boosting, Logistic Regression, Support Vector Machine (SVC), K-Nearest Neighbors, XGBoost, LightGBM, CatBoost, and Extra Trees\u0026mdash;and assigns optimal weights according to cross-validated accuracy scores. This ensemble framework was chosen for its robustness against overfitting and its ability to model complex nonlinear interactions across ecological variables.\u003c/p\u003e\u003cp\u003ePrior to modeling, data preprocessing included missing data treatment, outlier detection, and dimensionality reduction. Variables with more than 20% missingness were excluded. For the remaining variables, missing values were imputed using proximity-based imputation\u0026mdash;the mean for continuous variables and the mode for categorical variables, based on the most similar observations (Breiman, \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2001\u003c/span\u003e). Although more computationally intensive than simple imputation, this method effectively preserves nonlinear associations among predictors (Stekhoven \u0026amp; B\u0026uuml;hlmann, \u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e2012\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eTo mitigate multicollinearity and computational burden, Principal Component Analysis (PCA) was performed for dimensionality reduction. The dataset was then randomly partitioned into training (70%) and testing (30%) subsets. To address class imbalance across latent profiles, the Synthetic Minority Oversampling Technique (SMOTE) was applied to the training data (Mishra \u0026amp; Singh, \u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Model performance was evaluated using the Area Under the Receiver Operating Characteristic Curve (AUC) metric. The final CAWPE model achieved an AUC of 0.8788, indicating high predictive accuracy. Finally, partial dependence plots (PDPs) were generated to visualize the marginal effects of the top 20 most influential predictors on latent class membership, enhancing the interpretability of nonlinear relationships among ecological factors.\u003c/p\u003e\u003c/div\u003e"},{"header":"4 Results","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e4.1 Classification of Latent Profiles of Adolescents\u0026rsquo; Socio-Emotional Competence\u003c/h2\u003e\u003cp\u003eLatent Profile Analysis (LPA) was conducted to identify distinct socio-emotional profiles among adolescents using five standardized indicators\u0026mdash;creativity, perseverance, optimism, sociability, and empathy\u0026mdash;derived from the OECD SSES framework. Model fit indices for the two- to six-profile solutions are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Although the overall fit improved with additional classes (as indicated by reductions in AIC, BIC, and SA-BIC values), the rate of improvement diminished after the three-profile solution (Jedidi et al., \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e1997\u003c/span\u003e). Although the entropy value for the three-profile solution (0.748) fell slightly below the conventional threshold of 0.80 typically used to indicate strong classification quality (Clark \u0026amp; Muth\u0026eacute;n, 2009), it was deemed acceptable in light of the model's clear substantive interpretability. Moreover, while the Lo-Mendell-Rubin Likelihood Ratio Test (LMR-LRT) supported a five-profile solution (p\u0026thinsp;\u0026lt;\u0026thinsp;.001 for \u003cem\u003ek\u003c/em\u003e\u0026thinsp;=\u0026thinsp;5), solutions with four or more profiles produced classes with very small subgroup sizes (i.e., less than 5% of the sample), thereby limiting both interpretive stability and practical applicability (Jung \u0026amp; Wickrama, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e2008\u003c/span\u003e). Consequently, the three-profile model was retained as the most parsimonious and theoretically coherent representation of the data.\u003c/p\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. presents the standardized mean scores across the five indicators for each latent group. The first group (32.4%, n\u0026thinsp;=\u0026thinsp;736), labeled \u0026lsquo;Cognitively Engaged but Emotionally Detached\u0026rsquo;, exhibited moderate levels of cognitive and behavioral engagement but notably low cognitive empathy (Z = \u0026minus;\u0026thinsp;1.25), suggesting strong cognitive focus yet limited emotional responsiveness. The second group (47.5%, n\u0026thinsp;=\u0026thinsp;1,081), labeled \u0026lsquo;Emotionally Sensitive but Socially Withdrawn\u0026rsquo;, showed relatively high empathy (Z\u0026thinsp;\u0026gt;\u0026thinsp;0.6) but lower sociability (Z = \u0026minus;\u0026thinsp;0.4), reflecting emotional receptiveness combined with introverted and avoidant social tendencies. The third group (20.1%, n\u0026thinsp;=\u0026thinsp;458), labeled \u0026lsquo;Holistically Competent and Well-Adjusted\u0026rsquo;, exhibited consistently high scores (Z\u0026thinsp;\u0026gt;\u0026thinsp;0.5) across all cognitive, emotional, and social domains, representing a balanced and well-integrated profile of competence. The results reveal substantial heterogeneity in adolescents\u0026rsquo; socio-emotional development. Notably, discrepancies between emotional understanding (empathy) and behavioral enactment (sociability and perseverance) across groups indicate that empathic awareness does not necessarily translate into social adaptability or proactive engagement. This finding underscores the importance of differentiated intervention strategies tailored to each subgroup\u0026rsquo;s specific strengths and vulnerabilities.\u003c/p\u003e\u003cp\u003eAlthough cognitive empathy played a pivotal role in distinguishing between the profiles, classification was not determined by this indicator alone. As shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e, meaningful differentiation also arose from variation in creativity, perseverance, and sociability. For example, the \u0026lsquo;Holistically Competent and Well-Adjusted\u0026rsquo; group maintained uniformly high scores across all five indicators, whereas the \u0026lsquo;Emotionally Sensitive but Socially Withdrawn\u0026rsquo; Group displayed a marked imbalance between affective and behavioral domains. While the single-item nature of the empathy measure warrants interpretive caution, the classification patterns were multidimensionally informed, not dependent on a single construct. Furthermore, variability in sociability (range = \u0026minus;\u0026thinsp;0.4 to +\u0026thinsp;0.63) and perseverance (range = \u0026minus;\u0026thinsp;0.43 to +\u0026thinsp;0.73) contributed meaningfully to subgroup differentiation, reinforcing that socio-emotional competence reflects a complex interplay of emotional, cognitive, and behavioral components rather than a linear continuum of skill development.\u003c/p\u003e\u003cp\u003e\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\u003eFit indices and classification accuracy for latent profile models (2\u0026ndash;6 Classes)\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=\"left\" 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\" colspan=\"2\" morerows=\"1\" nameend=\"c2\" namest=\"c1\" rowspan=\"2\"\u003e\u003cp\u003eCriterion\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"4\" nameend=\"c6\" namest=\"c3\"\u003e\u003cp\u003eNumber of Latent Profiles\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u0026nbsp;\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003e2 classes\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e3 classes\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e4 classes\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003e5 classes\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003e6 classes\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e\u003cp\u003eInformation Indices\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e31529.214\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e31007.603\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e30805.841\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e30686.937\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e30637.914\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eBIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e31620.890\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e31133.657\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e30966.274\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e30881.748\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e30867.103\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSA-BIC\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e31570.055\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e31063.759\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e30877.313\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e30773.724\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e30740.016\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eQuality of Classification\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eEntropy\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.881\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.748\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0.754\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0.745\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.753\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eModel Comparison\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAdjusted LMR (p)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.079\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eB-LRT(p)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e0.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;.001\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0.083\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"5\" rowspan=\"6\"\u003e\u003cp\u003eClassification Rate\u003c/p\u003e\u003cp\u003e(Number, %)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e724(31.8)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e736(32.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e719(31.6)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e617(27.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e420 (18.5)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e1,551(68.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e1,081(47.5)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e918(40.4)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e458(20.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e595 (26.2)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e458(20.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e527(23.2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e138(6.1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e963(42.3)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eClass4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\" morerows=\"2\" rowspan=\"3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\" morerows=\"2\" rowspan=\"3\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e111(4.9)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e948(41.7)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e156(6.9)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eClass5\u003c/p\u003e\u003cp\u003eClass6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e114(5.0)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e48 (2.1)\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e93 (4.1)\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\u003cstrong\u003eNote\u003c/strong\u003e\u003cp\u003eWhile empathy showed the largest difference between profiles, meaningful variation was also observed in sociability (range: \u0026minus;\u0026thinsp;0.4 to +\u0026thinsp;0.63), perseverance (\u0026ndash;0.43 to +\u0026thinsp;0.73), and creativity (\u0026ndash;0.11 to +\u0026thinsp;0.78), indicating that classification was multidimensional.\u003c/p\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec10\" class=\"Section2\"\u003e\u003ch2\u003e4.2 Predictive Factors of Socio-Emotional Competence Subgroup Classification\u003c/h2\u003e\u003cp\u003eThe Cross-validation Accuracy Weighted Probabilistic Ensemble (CAWPE) model achieved strong predictive performance (AUC\u0026thinsp;=\u0026thinsp;0.8788), confirming the robustness of ecological predictors. Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e lists the twenty most influential variables, and Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. illustrates the marginal effects of the major predictors on subgroup membership.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cb\u003eIndividual-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eBehavioral and psychological characteristics were key differentiators among socio-emotional profiles. Adequate sleep (SL, SL1) and balanced physical activity (YTIM1J01w4) increased the likelihood of belonging to the \u0026lsquo;Holistically Competent and Well-Adjusted\u0026rsquo; group, while excessive gaming (YTIM1L02w4) and unstructured smartphone use (YTIM1K01w4) predicted membership in the \u0026lsquo;Cognitively Engaged but Emotionally Detached\u0026rsquo; group. Notably, purposeful digital engagement, such as information seeking (YMDA1B11w4), rather than mere usage volume, was more beneficial.\u003c/p\u003e\u003cp\u003ePsychologically, adolescents with a strong sense of personal strengths (\u0026ldquo;I have many strengths,\u0026rdquo; YPSY3A03w4) were more likely to belong to the \u0026lsquo;Holistically Competent\u0026rsquo; group, whereas those reporting low self-affirmation (\u0026ldquo;I have few things to be proud of,\u0026rdquo; YPSY3A05w4) tended to cluster in the \u0026lsquo;Emotionally Sensitive but Socially Withdrawn\u0026rsquo; group.\u003c/p\u003e\u003cp\u003e\u003cb\u003eFamily-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eFamily-level factors emerged as the most critical predictors, with parental cognitive empathy (PRME1Ascw4) identified as the single most dominant factor by a substantial margin. Its importance index (69.70) was more than three times higher than the second-ranked predictor (Time spent with parents on weekdays, 18.17), suggesting that a parent's capacity for empathy is the primary ecological foundation shaping adolescent socio-emotional profiles. Higher parental empathy substantially increased the likelihood of adolescents belonging to the \u0026lsquo;Holistically Competent and Well-Adjusted\u0026rsquo; group while reducing the likelihood of belonging to the \u0026lsquo;Cognitively Engaged but Emotionally Detached\u0026rsquo; group.\u003c/p\u003e\u003cp\u003eOther family-level variables also showed significant predictive power. Regular parent\u0026ndash;child interaction, especially on weekends (SPEND2, Importance\u0026thinsp;=\u0026thinsp;17.71), increased the probability of \u0026lsquo;Holistic Competence\u0026rsquo;. Similarly, parental creative disposition (PPSY4Ascw4, Importance\u0026thinsp;=\u0026thinsp;11.18) and higher household income (PINCOMEw4, Importance\u0026thinsp;=\u0026thinsp;12.69) were important predictors that helped differentiate the \u0026lsquo;Holistically Competent\u0026rsquo; group from the two less-adjusted profiles.\u003c/p\u003e\u003cp\u003e\u003cb\u003eSchool-Level Factors\u003c/b\u003e\u003c/p\u003e\u003cp\u003eSchool-related predictors also significantly distinguished the groups. Positive teacher recognition (e.g., \u0026ldquo;The teacher considers me to be a smart student,\u0026rdquo; YEDU3A14w4) increased the probability of belonging to the 'Holistically Competent' group. A moderate willingness to seek teacher counseling (YEDU3A01w4) was optimal, differentiating balanced students from those overly dependent (Emotionally Sensitive) or emotionally distant (Emotionally Detached).\u003c/p\u003e\u003cp\u003e Similarly, the frequency of conversations with a career counselor (YFUR2A06w4) showed a U-shaped relationship, with both very low and very high counseling frequencies linked to higher competence. Moderate peer interaction\u0026mdash;during weekdays (YTIM1N02w4) and weekends (YTIM1N01w4) \u0026mdash;was associated with 'Holistic Competence', while too little interaction aligned with 'Social Withdrawal' and excessive interaction with 'Emotional Detachment'. Moreover, higher agreement with \u0026ldquo;I do not want to be close to other students\u0026rdquo; (YEDU2A12w4) predicted the two less adjusted profiles.\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\u003eKey predictors of adolescents\u0026rsquo; socio-emotional competence classification\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\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\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eRank\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eVariable\u003c/p\u003e\u003cp\u003eName\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDomain\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eQuestionnaire\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eImportance Index\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePRME1Ascw4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParental Competencies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eParent\u0026rsquo;s cognitive empathy score\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e69.70\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSPEND1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParent-Child Relationship\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent with parents on weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e18.17\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSPEND2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParent-Child Relationship\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent with parents on weekends\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e17.71\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSL1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eSleep duration on weekends\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e14.50\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1L02w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent playing with a computer on weekends\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e14.17\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1N01w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent playing with friends outside of weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e13.81\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e7\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYPSY3A03w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePhysical and Psychological Characteristics\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eI perceive myself as having many strengths.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e13.46\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e8\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYFUR2A06w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eFrequency of career-related discussions with a school counselor\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e13.05\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e9\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1M01w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent watching TV and playing on weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.97\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e10\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSL\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003esleep duration on weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.80\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e11\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePINCOMEw4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eHome Background\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eMonthly average household income\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.69\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e12\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1N02w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent with friends outside of weekends\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e12.16\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYEDU2A12w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eI have no intention of becoming close with other children.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.60\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYPSY3A05w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003ePhysical and Psychological Characteristics\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eI feel that I have few things to be proud of.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.50\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYMDA1B11w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eFrequency of smartphone use by purpose: view document\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.48\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e16\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1K01w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent using a smartphone on weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.41\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e17\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYEDU3A14w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eThe teacher considers me to be a smart student.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.33\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e18\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYTIM1J01w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eIndividual Backgrounds and Living Conditions\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTime spent on exercise and physical activity on weekdays\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.31\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e19\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePPSY4Ascw4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eParental Competencies\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eParent\u0026rsquo;s creativity personality score\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.18\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e20\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eYEDU3A01w4\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eRelationship and School Life\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003eWhen I am struggling with studying or other issues, I would like to consult with the teacher first.\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e11.17\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"5 Discussion","content":"\u003cp\u003eThis study provides empirical evidence that adolescents\u0026rsquo; socio-emotional competence is not fixed or developed in a linear trajectory but rather unfolds as a dynamic process shaped by the interplay among cognitive, emotional, and social components within broader ecological contexts. The three latent profiles identified through Latent Profile Analysis (LPA)\u0026mdash;\u0026ldquo;Holistically Competent,\u0026rdquo; \u0026ldquo;Emotionally Sensitive but Socially Withdrawn,\u0026rdquo; and \u0026ldquo;Cognitively Engaged but Emotionally Detached\u0026rdquo;\u0026mdash;highlight the heterogeneity in socio-emotional development. Notably, the observed \u003cem\u003ediscrepancy\u003c/em\u003e between emotional understanding (e.g., empathy) and behavioral enactment (e.g., sociability) challenges the prevailing notion that socio-emotional competence can be captured along a single, unified dimension. The identification of these distinct profiles underscores the methodological strength of the present study. For instance, adolescents in the \u0026ldquo;Cognitively Engaged but Emotionally Detached\u0026rdquo; group scored at average levels on self-report measures but demonstrated significantly lower performance on the Reading the Mind in the Eyes Test (RMET), an objective measure of cognitive empathy. Had this study relied solely on self-report instruments, such hidden deficits in emotional functioning may have remained undetected. The integration of both self-report and performance-based assessments was therefore essential to uncovering the complex\u0026mdash;and at times contradictory\u0026mdash;nature of socio-emotional competence.\u003c/p\u003e\u003cp\u003eThe results of the machine learning\u0026ndash;based predictive analysis further corroborate Bronfenbrenner\u0026rsquo;s ecological model, revealing that variables within the adolescent\u0026rsquo;s microsystem exert the greatest influence on socio-emotional development. Among all ecological predictors, parental cognitive empathy emerged as the most dominant and consistent predictor, surpassing all other variables in predictive strength. This finding represents one of the study\u0026rsquo;s most novel and policy-relevant contributions, suggesting that parents' empathic dispositions form a foundational ecological bedrock for adolescents' socio-emotional growth. Theoretically, this study contributes by providing a practical and scalable approach to empirically operationalizing the OECD\u0026rsquo;s Study on Social and Emotional Skills (SSES) framework using nationally representative panel data. In doing so, it offers a methodological model for translating complex theoretical constructs into policy-relevant indicators, particularly within the constraints of applied research settings. This aligns closely with the core objectives of the field of child indicator research\u0026mdash;namely, to inform child well-being policy through rigorous, multidimensional measurement.\u003c/p\u003e\u003cp\u003eThe findings yield several actionable implications for policy interventions aimed at enhancing adolescents\u0026rsquo; socio-emotional skills. First, a shift in intervention focus from students to parents is warranted. The fact that \u003cem\u003eparental cognitive empathy\u003c/em\u003e was over 3.8 times more predictive than any school- or student-level factor suggests that investing limited policy resources in parent-targeted empathy training and counseling programs may be the most cost-effective strategy. Second, a one-size-fits-all approach to intervention is unlikely to be effective. The discovery of three qualitatively distinct socio-emotional profiles points to divergent needs across subgroups. For example, adolescents in the \u0026ldquo;Emotionally Sensitive but Socially Withdrawn\u0026rdquo; profile may not benefit from generic empathy-building curricula, but instead require targeted social skills training to help them translate emotional sensitivity into behavioral expression.\u003c/p\u003e\u003cp\u003eDespite its contributions, this study is not without limitations. First, the use of cross-sectional data restricts the ability to draw causal inferences between parental empathy and adolescents\u0026rsquo; socio-emotional competencies. Second, due to structural constraints of large-scale panel data, certain subdomains (e.g., perseverance, sociability) were assessed using a limited number of items\u0026mdash;a trade-off necessary to balance practical feasibility with international comparability. Future research should address these limitations by employing longitudinal designs to examine the developmental trajectories of socio-emotional profiles and test causal pathways involving ecological predictors. Additionally, expanding the scope to include macrosystem-level factors such as educational policy and cultural values will be essential for building a more comprehensive, multilayered model of socio-emotional development.\u003c/p\u003e"},{"header":"6 Conclusion","content":"\u003cp\u003eGuided by the OECD Study on Social and Emotional Skills (SSES) framework, this study aimed to identify distinct socio-emotional competence (SEC) profiles among Korean adolescents and to uncover their key ecological predictors using machine learning techniques. The findings revealed that adolescent SEC does not follow a single, uniform trajectory; rather, it manifests across three qualitatively distinct subgroups\u0026mdash;\u0026ldquo;Holistically Competent and Well-Adjusted,\u0026rdquo; \u0026ldquo;Emotionally Sensitive but Socially Withdrawn,\u0026rdquo; and \u0026ldquo;Cognitively Engaged but Emotionally Detached.\u0026rdquo; These results highlight the non-linear and heterogeneous nature of competence development during adolescence and underscore the need for differentiated policy approaches tailored to the specific strengths and vulnerabilities of each subgroup.\u003c/p\u003e\u003cp\u003eAmong the study\u0026rsquo;s most significant contributions is the empirical identification of parental cognitive empathy as the single most powerful and dominant ecological predictor of adolescents\u0026rsquo; socio-emotional competence. Among dozens of individual-, family-, and school-level variables, this factor surpassed all others in predictive strength. This finding has critical implications for public policy, suggesting that fostering empathic capacities in parents may serve as the most effective leverage point for promoting healthy adolescent development.\u003c/p\u003e\u003cp\u003eIn conclusion, this study offers both theoretical and methodological contributions to the field of child indicator research, while simultaneously providing robust empirical evidence to inform practice. By demonstrating that parental empathy is a central ecological driver of adolescent socio-emotional outcomes, the findings establish a clear and actionable direction for designing impactful and resource-efficient interventions.\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eThe sole author was responsible for the conception and design of the study, data analysis and interpretation, as well as the drafting and revision of the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBlakemore, S., \u0026amp; Mills, K. (2014). Is adolescence a sensitive period for sociocultural processing?. \u003cem\u003eAnnual Review of Psychology, 65\u003c/em\u003e, 187-207.\u003c/li\u003e\n\u003cli\u003eBreiman, L. (2001). Random forests. \u003cem\u003eMachine learning, 45\u003c/em\u003e(1), 5-32.\u003c/li\u003e\n\u003cli\u003eBronfenbrenner, U. (1994). \u003cem\u003eEcological models of human development\u003c/em\u003e. In International Encyclopedia of Education. Oxford: Elsevier.\u003c/li\u003e\n\u003cli\u003eBrown, L., Green, K., \u0026amp; Taylor, D. (2021). Coping with COVID-19: Adolescent stress and coping strategies. \u003cem\u003eJournal of Youth and Adolescence, 50\u003c/em\u003e(5), 987-1001.\u003c/li\u003e\n\u003cli\u003eBurke, J. D., \u0026amp; Loeber, R. (2016). Mechanisms of behavioral and affective treatment outcomes in a cognitive behavioral intervention for boys. \u003cem\u003eJournal of Abnormal Child Psychology, 44\u003c/em\u003e(1), 179\u0026ndash;189.\u003c/li\u003e\n\u003cli\u003eCasey, B. (2015). Beyond simple models of self-control to circuit-based accounts of adolescent behavior. \u003cem\u003eAnnual Review of Psychology, 66\u003c/em\u003e, 295-319.\u003c/li\u003e\n\u003cli\u003eChernyshenko, O., M. Kankara\u0026scaron; and F. Drasgow (2018). Social and emotional skills for student success and well-being: Conceptual framework for the OECD study on social and emotional skills. OECD Education Working Papers, 173, OECD Publishing: Paris. \u003c/li\u003e\n\u003cli\u003eCollaborative for academic, social, and emotional learning. (2015). 2015 CASEL guide: Effective social and emotional learning programs \u0026ndash; Middle and high school edition. Chicago, IL: Author.\u003c/li\u003e\n\u003cli\u003eDatu, J., \u0026amp; Restubog, S. (2020). The emotional pay-off of staying gritty: linking grit with social-emotional learning and emotional well-being. \u003cem\u003eBritish Journal of Guidance \u0026amp; Counselling, 48\u003c/em\u003e(5), 697\u0026ndash;708. \u003c/li\u003e\n\u003cli\u003eDurlak, J., Weissberg, R., Dymnicki, A., Taylor, R., \u0026amp; Schellinger, K. (2011). The impact of enhancing students\u0026rsquo; social and emotional learning: A meta-analysis of school-based universal interventions. \u003cem\u003eChild Development, 82\u003c/em\u003e(1), 405-432.\u003c/li\u003e\n\u003cli\u003eGlick, G., \u0026amp; Rose, A. (2011): Prospective associations between friendship adjustment and social strategies: Friendship as a context for building social skills. \u003cem\u003eDevelopmental Psychology, 47\u003c/em\u003e(4), 1117-1132.\u003c/li\u003e\n\u003cli\u003eHa, H., Choi, Y., Jeong, E., Jeong, Y., \u0026amp; Han, J. (2017). Korean Children and Youth Panel Survey VIII: Project report. National Youth Policy Institute.\u003c/li\u003e\n\u003cli\u003eHan, S., Kim, H., Seol, I., Lim, Y., \u0026amp; Cho, A. (2014). Adolescent Psychology (2nd ed.). Gyeonggi: Education Science Press.\u003c/li\u003e\n\u003cli\u003eHuhtala, M., Korja, R., Lehtonen, L., Haataja, L., Lapinleimu, H., Rautava, P., \u0026amp; PIPARI Study Group. (2014). Associations between parental psychological well-being and socio-emotional development in 5-year-old preterm children. \u003cem\u003eEarly Human Development, 90\u003c/em\u003e(3), 119-124. \u003c/li\u003e\n\u003cli\u003eJedidi, K., Jagpal, H. S., \u0026amp; DeSarbo, W. (1997). Finite-mixture structural equation models for response-based segmentation and unobserved heterogeneity. \u003cem\u003eMarketing Science, 16\u003c/em\u003e(1), 39-59.\u003c/li\u003e\n\u003cli\u003eJones, G., White, R., \u0026amp; Smith, P. (2022). Effects of remote learning on social-emotional development in adolescents. \u003cem\u003eEducational Psychology Review, 34\u003c/em\u003e(1), 123-145.\u003c/li\u003e\n\u003cli\u003eJung, T., \u0026amp; Wickrama, K. A. (2008). An introduction to latent class growth analysis and growth mixture modeling. \u003cem\u003eSocial and personality psychology compass, 2\u003c/em\u003e(1), 302-317.\u003c/li\u003e\n\u003cli\u003eKim, H., Kim, M., Lee, S., Yang, H., Kim, J., Kim, M., Kim, J., \u0026amp; Cho, Y. (2020). OECD ESP Social and Emotional Skills Survey International Collaborative Study (IV): Analysis of Main Survey Results. Chungbuk: Korea Educational Development Institute.\u003c/li\u003e\n\u003cli\u003eKim, K., Kim, W., Choi, I., Kim, M., Kim, K., Park, J., Kim, S., Park, J., Do, S., \u0026amp; Jang, H. (2020). Development of student competency assessment tool for the 2020 Seoul education longitudinal study. Korea Institute for Curriculum and Evaluation. Research Report.\u003c/li\u003e\n\u003cli\u003eMartinez-Yarza, N., Santib\u0026aacute;\u0026ntilde;ez, R., \u0026amp; Solabarrieta, J. (2023). A Systematic Review of Instruments Measuring Social and Emotional Skills in School-Aged Children and Adolescents. \u003cem\u003eChild Indicators Research, 16\u003c/em\u003e(4), 1475\u0026ndash;1502.\u003c/li\u003e\n\u003cli\u003eMishra, N., \u0026amp; Singh, P. (2021). Feature construction and smote-based imbalance handling for multi-label learning. \u003cem\u003eInformation Sciences, 563\u003c/em\u003e, 342-357.\u003c/li\u003e\n\u003cli\u003eNapolitano, C. M., Greenberg, M. T., \u0026amp; Weissberg, R. P. (2021). Toward a Developmental Framework of Social-Emotional Competence for Youth. \u003cem\u003eAmerican Psychologist, 76\u003c/em\u003e(6), 878\u0026ndash;893.\u003c/li\u003e\n\u003cli\u003eOECD (2015). Skills for social progress: The power of social and emotional skills. OECD Publishing.\u003c/li\u003e\n\u003cli\u003eOrth, U., Robins, R., \u0026amp; Widaman, K. (2012). Life-span development of self-esteem and its effects on important life outcomes. \u003cem\u003eJournal of Personality and Social Psychology, 102\u003c/em\u003e(6), 1271-1288.\u003c/li\u003e\n\u003cli\u003ePeetz, H. K., Lansu, T. A., Hoekstra, N. A., van den Berg, Y. H., Burk, W. J., \u0026amp; Mainhard, M. T. (2025). Assessing youth\u0026apos;s internal and external attributions to negative peer interactions and victimization\u0026mdash;development of the Causal Attributions for Peer Experiences (CAPE) scale. \u003cem\u003eJournal of Research on Adolescence\u003c/em\u003e, \u003cem\u003e35\u003c/em\u003e(2), e70037.\u003c/li\u003e\n\u003cli\u003ePortela-Pino, I., Alvari\u0026ntilde;as-Villaverde, M., \u0026amp; Pino-Juste, M. (2021). Socio-emotional skills in adolescence. Influence of personal and extracurricular variables. \u003cem\u003eInternational Journal of Environmental Research and Public Health, 18\u003c/em\u003e(9), 4811.\u003c/li\u003e\n\u003cli\u003eRoorda, D., Koomen, H., Spilt, J., \u0026amp; Oort, F. (2011). The influence of affective teacher\u0026ndash;student relationships on students\u0026rsquo; school engagement and achievement: A meta-analytic approach. \u003cem\u003eReview of Educational Research, 81\u003c/em\u003e(4), 493-529.\u003c/li\u003e\n\u003cli\u003eRoss, K. M., Kim, H., Tolan, P. H., \u0026amp; Jennings, P. A. (2019). An exploration of normative social and emotional skill growth trajectories during adolescence. \u003cem\u003eJournal of Applied Developmental Psychology, 62\u003c/em\u003e, 151\u0026ndash;162.\u003c/li\u003e\n\u003cli\u003eSalmela-Aro, K., \u0026amp; Upadyaya, K. (2020). School engagement and school burnout profiles during high school\u0026ndash;The role of socio-emotional skills. \u003cem\u003eEuropean Journal of Developmental Psychology, 17\u003c/em\u003e(6), 943-964. \u003c/li\u003e\n\u003cli\u003eSavickas, M. (2013). Career construction theory and practice. Career development and counseling. \u003cem\u003ePutting theory and research to work, 2,\u003c/em\u003e 144-180.\u003c/li\u003e\n\u003cli\u003eSklad, M., Diekstra, R., Ritter, M. D., Ben, J., \u0026amp; Gravesteijn, C. (2012). Effectiveness of school‐based universal social, emotional, and behavioral programs: Do they enhance students\u0026rsquo; development in the area of skill, behavior, and adjustment?. \u003cem\u003ePsychology in the Schools, 49\u003c/em\u003e(9), 892-909.\u003c/li\u003e\n\u003cli\u003eSmith, A., Brown, C., \u0026amp; Jones, E. (2021). The impact of the COVID-19 pandemic on adolescent mental health. \u003cem\u003eJournal of Adolescent Health, 68\u003c/em\u003e(3), 456-463.\u003c/li\u003e\n\u003cli\u003eStekhoven, D. J., \u0026amp; B\u0026uuml;hlmann, P. (2012). MissForest\u0026mdash;non-parametric missing value imputation for mixed-type data. \u003cem\u003eBioinformatics, 28\u003c/em\u003e(1), 112-118.\u003c/li\u003e\n\u003cli\u003eThompson, R. B., Corsello, M., McReynolds, S., \u0026amp; Conklin-Powers, B. (2013). A Longitudinal study of family socioeconomic status (SES) variables as predictors of socio-emotional resilience among mentored youth. \u003cem\u003eMentoring \u0026amp; Tutoring: Partnership in Learning, 21\u003c/em\u003e(4), 378\u0026ndash;391. \u003c/li\u003e\n\u003cli\u003eYu, M., \u0026amp; Lee, K. (2023). Study on the implementation of the 2023 Convention on the Rights of the Child: Basic analysis report on the status of children\u0026rsquo;s and adolescents\u0026rsquo; rights in Korea. National Youth Policy Institute.\u003c/li\u003e\n\u003cli\u003eWoo, Y., Lee, J., Lee, I., Hong, W., Choi, J., Jung, E., \u0026amp; Kwon, Y. (2022). Development of diagnostic tools and guidance/support materials for socio-emotional competence. Korea Institute for Curriculum and Evaluation, Research Report.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":true,"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":"Social and emotional skills, Child indicators, OECD SSES framework, Latent profile analysis, Machine learning","lastPublishedDoi":"10.21203/rs.3.rs-7891642/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7891642/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study investigates the multidimensional structure and ecological predictors of Korean adolescents\u0026rsquo; socio-emotional skills using a person-centered, data-driven approach grounded in the OECD Social and Emotional Skills (SSES) framework. Utilizing nationally representative data from the Korean Children and Youth Panel Survey (KCYPS; N\u0026thinsp;=\u0026thinsp;2,275), latent profile analysis identified three distinct subgroups: (1) Holistically Competent and Well-Adjusted (20.1%), (2) Emotionally Sensitive but Socially Withdrawn (47.5%), and (3) Cognitively Engaged but Emotionally Detached (32.4%). To determine ecological predictors of subgroup membership, a stacked ensemble machine learning model (CAWPE) was employed, integrating individual, familial, and school-level variables. The most influential predictors included parental cognitive empathy and creativity, adolescents\u0026rsquo; self-esteem, peer and teacher relationships, household income, and engagement in leisure and career-related activities. Findings reveal that socio-emotional development in early adolescence is not a linear process but a multidimensional interplay of psychological, relational, and contextual factors within proximal ecological systems. By empirically operationalizing the OECD SSES framework using large-scale national data, this study provides a novel cross-level validation of socio-emotional skill indicators and demonstrates how person-centered and predictive analytics can inform evidence-based, developmentally targeted interventions and policy design.\u003c/p\u003e","manuscriptTitle":"Ecological and Predictive Indicators of Social and Emotional Skills among Korean Adolescents: A Person-Centered Analysis within the OECD SSES Framework","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-13 08:15:02","doi":"10.21203/rs.3.rs-7891642/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":"975afc1f-6572-4c78-98fa-7b23132bd687","owner":[],"postedDate":"November 13th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-16T16:07:55+00:00","versionOfRecord":{"articleIdentity":"rs-7891642","link":"https://doi.org/10.1007/s12187-026-10355-w","journal":{"identity":"child-indicators-research","isVorOnly":false,"title":"Child Indicators Research"},"publishedOn":"2026-03-13 15:58:27","publishedOnDateReadable":"March 13th, 2026"},"versionCreatedAt":"2025-11-13 08:15:02","video":"","vorDoi":"10.1007/s12187-026-10355-w","vorDoiUrl":"https://doi.org/10.1007/s12187-026-10355-w","workflowStages":[]},"version":"v1","identity":"rs-7891642","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7891642","identity":"rs-7891642","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}