Journal of Adolescent and Youth Psychological Studies (JAYPS)

Objective: The objective of this study was to develop and validate an interpretable deep learning model for predicting adolescents’ online social anxiety and to identify the most influential psychological and digital behavioral features contributing to its development.

Methods and Materials: A cross-sectional predictive modeling design was employed with a sample of 523 adolescents aged 13–18 years recruited from secondary schools in Austria. Participants completed standardized measures assessing online social anxiety, general social anxiety, self-esteem, perceived social support, emotion regulation, and digital behavior patterns. Data were analyzed using a deep neural network with multiple hidden layers and regularization techniques. Model performance was evaluated using RMSE, MAE, R², and Pearson correlation. Interpretability was achieved through advanced feature saliency mapping, including SHAP values and gradient-based attribution methods, enabling identification of both global and individual-level predictors of online social anxiety.

Findings: The deep learning model demonstrated strong predictive performance, explaining 81% of the variance in online social anxiety (R² = .81) with high correspondence between predicted and observed scores (r = .90). Feature saliency analysis revealed that online social comparison exerted the strongest positive influence on anxiety predictions, followed by low self-esteem, emotion dysregulation, general social anxiety, and daily social media use. Perceived social support displayed a robust negative contribution, functioning as a protective factor. Subgroup analysis indicated that adolescents with the highest anxiety levels exhibited intensified contributions from social comparison, emotional instability, excessive media engagement, and negative feedback sensitivity.

Conclusion: Interpretable deep learning provides a powerful and transparent framework for understanding the complex psychological mechanisms underlying adolescents’ online social anxiety, offering critical insights for early identification and targeted intervention.

Objective: The objective of this study was to develop and interpret an explainable artificial intelligence model of youth hope by quantifying the individual and interactive contributions of optimism, goal orientation, and family support.

Methods and Materials: This cross-sectional study was conducted with 653 adolescents and emerging adults aged 15–24 from educational and community institutions in Georgia, United States. Participants completed standardized measures of hope, optimism, goal orientation, and perceived family support. Machine learning models including Random Forest, Gradient Boosting, and XGBoost were trained to predict hope, with model performance evaluated using cross-validation and error metrics. Explainability was achieved through SHAP and permutation-based feature attribution methods, enabling identification of global and individual predictor effects and nonlinear interactions.

Findings: The XGBoost model demonstrated the strongest predictive performance, explaining 84% of the variance in youth hope (R² = .84, RMSE = 2.19, MAE = 1.68). Feature attribution analyses indicated that family support contributed the largest proportion of predictive influence (38.6%), followed by optimism (32.1%) and goal orientation (25.3%). Significant interaction effects were observed between optimism and family support, as well as between goal orientation and family support, amplifying their combined impact on hope.

Conclusion: Youth hope is best explained as a multilevel construct arising from the integrated influence of motivational beliefs and family relational processes. Explainable AI offers a powerful methodological framework for advancing theoretical understanding and guiding personalized intervention strategies.

Objective: The objective of this study was to identify and quantify the relative and interactive contributions of psychosocial, relational, cultural, and individual predictors of adolescents’ identity development using feature sensitivity mapping within an interpretable machine learning framework.

Methods and Materials: This cross-sectional study was conducted among 1,032 adolescents aged 14–18 years recruited from public and private secondary schools in three major metropolitan regions of Brazil using multi-stage cluster sampling. Participants completed validated measures assessing identity development dimensions, parental autonomy support, peer attachment quality, emotional self-regulation, future orientation, academic self-efficacy, social competence, and demographic characteristics. Data were analyzed using an ensemble of gradient boosting and random forest models with ten-fold cross-validation and Bayesian hyperparameter optimization. Model interpretability was achieved through SHAP-based feature sensitivity mapping, partial dependence plots, and accumulated local effects analysis to examine nonlinear and interactive predictor effects.

Findings: The machine learning models explained substantial variance in identity outcomes, with the highest performance observed for identity commitment (R² = .61), followed by exploration in breadth (R² = .57), exploration in depth (R² = .55), and ruminative exploration (R² = .49). Parental autonomy support, peer attachment quality, emotional self-regulation, future orientation, academic self-efficacy, and social competence emerged as the most influential predictors across identity dimensions. Emotional self-regulation demonstrated the strongest protective effect against ruminative exploration, while parental autonomy support and peer attachment quality exhibited the highest positive contributions to identity commitment.

Conclusion: Adolescents’ identity development operates as a nonlinear, interactive system in which relational security, self-regulation, and motivational resources play central roles, and interpretable machine learning provides a powerful framework for advancing developmental theory and intervention design.

Objective: This study aimed to develop an interpretable machine learning model to examine the predictive roles of moral disengagement and achievement goal orientations in academic cheating among high school adolescents.

Methods and Materials: The study employed a cross-sectional correlational design with a predictive analytics framework and was conducted among 681 adolescents aged 14–18 years enrolled in public high schools in California. Participants completed standardized self-report measures assessing academic cheating behavior, moral disengagement, and achievement goal orientations, along with demographic information. Data were analyzed using an Extreme Gradient Boosting (XGBoost) regression model with five-fold cross-validation and Bayesian hyperparameter optimization. Model performance was evaluated using root mean squared error, mean absolute error, and explained variance. To ensure interpretability, Shapley Additive Explanations were applied to quantify the relative and local contributions of predictors, and partial dependence analyses were conducted to examine nonlinear and interactive effects.

Findings: The gradient boosting model demonstrated strong predictive performance, accounting for 56% of the variance in academic cheating. Moral disengagement emerged as the most influential predictor, followed by performance-avoidance and performance-approach goals. Mastery-approach goals exhibited a consistent negative association with cheating. The model identified nonlinear threshold effects for moral disengagement and significant interaction patterns between motivational orientations and moral cognition, indicating that performance-based goals amplified the impact of moral disengagement on cheating behavior.

Conclusion: The findings indicate that academic cheating in adolescence is primarily driven by cognitive moral mechanisms operating in conjunction with achievement-related motivational pressures. Interpretable machine learning offers a powerful framework for uncovering these complex psychological dynamics and provides actionable insights for the design of targeted educational interventions aimed at promoting academic integrity.

Objective: The objective of this study was to develop an explainable predictive model of youth life satisfaction by quantifying the individual and combined contributions of psychological capital components and social connectedness.

Methods and Materials: A cross-sectional correlational design was employed with a sample of 684 Moroccan adolescents and emerging adults recruited from secondary schools, vocational institutes, and universities. Participants completed standardized measures of life satisfaction, psychological capital, and social connectedness. Advanced machine learning models including Elastic Net Regression, Random Forest, and Gradient Boosting Machine were trained using ten-fold cross-validation. Explainable artificial intelligence techniques based on SHAP values and permutation feature importance were applied to interpret model predictions, identify dominant predictors, and examine nonlinear interactions among psychological and social variables.

Findings: The Gradient Boosting Machine demonstrated superior predictive performance (R² = .76, RMSE = 2.78, MAE = 2.19). Feature contribution analysis revealed that hope was the strongest predictor of life satisfaction, followed by social connectedness and self-efficacy. Optimism and resilience showed moderate but substantial contributions, whereas demographic variables such as age and gender exerted comparatively minor influence. Interaction effects indicated that high social connectedness amplified the positive effects of psychological capital components on predicted life satisfaction.

Conclusion: Youth life satisfaction is primarily driven by dynamic psychological and social resources rather than static demographic characteristics. Explainable machine learning provides a powerful framework for uncovering complex predictive structures and offers actionable insight for designing personalized youth well-being interventions focused on strengthening psychological capital and social integration.

Objective: The objective of this study was to develop and interpret an explainable artificial intelligence model to predict adolescents’ psychological adjustment from coping styles, social support quality, and life stress exposure in a representative sample of Iraqi secondary school students.

Methods and Materials: This cross-sectional study was conducted among 612 adolescents aged 14–18 years recruited from public secondary schools in Baghdad, Najaf, and Basra using multistage cluster sampling. Participants completed validated measures assessing psychological adjustment, coping styles, perceived social support quality, and life stress exposure. Data were preprocessed through normalization, missing-value imputation, and outlier screening. Machine learning models including random forest, extreme gradient boosting, and multilayer perceptron neural networks were trained using five-fold cross-validation. Model interpretability was achieved through SHapley Additive exPlanations, permutation feature importance, and partial dependence analyses.

Findings: The XGBoost model demonstrated the highest predictive performance (R² = 0.76, RMSE = 0.29, MAE = 0.22). Social support quality emerged as the strongest positive predictor of psychological adjustment, followed by problem-focused coping. Life stress exposure exerted a substantial negative effect. Avoidance coping significantly predicted poorer adjustment, whereas emotion-focused coping displayed nonlinear effects depending on stress levels. Interaction analysis revealed that high social support significantly buffered the adverse effects of life stress on psychological adjustment.

Conclusion: The findings demonstrate that adolescents’ psychological adjustment is governed by complex nonlinear interactions among coping strategies, social support, and stress exposure, and that explainable artificial intelligence offers a powerful framework for modeling these processes with high predictive accuracy and theoretical transparency.

Objective: This study aimed to examine the nonlinear and interactive effects of loneliness, family communication quality, and digital media dependency on depressive symptoms among Indonesian adolescents using neural network modeling.

Methods and Materials: A cross-sectional design was implemented with 684 secondary school students aged 14–18 years selected via multistage cluster sampling from urban regions of Indonesia. Participants completed validated measures assessing depressive symptoms, loneliness, family communication quality, and digital media dependency. Data were analyzed using a multilayer perceptron neural network constructed in Python with TensorFlow. The dataset was partitioned into training, validation, and test sets. Model performance was evaluated using root mean square error, mean absolute error, and coefficient of determination. Predictor contributions and interaction effects were interpreted using SHAP values and partial dependence analyses.

Findings: The neural network demonstrated high predictive accuracy (R² = .79 on the test set). Loneliness emerged as the strongest predictor of depressive symptoms, followed by digital media dependency and family communication quality. Significant nonlinear interaction effects were observed, indicating that the combination of high loneliness, poor family communication, and elevated digital media dependency produced the highest levels of depressive symptoms. Family communication quality exerted a strong buffering effect that attenuated the impact of loneliness and digital dependency on depression.

Objective: The objective of this study was to develop and validate an explainable artificial intelligence model for predicting academic burnout among Chinese high school students based on cognitive flexibility, perceived school climate, and online learning engagement.

Methods and Materials: This quantitative cross-sectional study was conducted with 1,042 high school students from public schools in three major urban regions of eastern China using multi-stage cluster sampling. Participants completed standardized measures of academic burnout, cognitive flexibility, school climate, and online learning engagement. Data were analyzed using an ensemble machine learning framework combining Random Forest, Gradient Boosting, and XGBoost algorithms. Model performance was evaluated via nested cross-validation. Explainable AI techniques including SHapley Additive exPlanations and Local Interpretable Model-agnostic Explanations were applied to ensure transparency and interpretability of predictions.

Findings: The ensemble model demonstrated strong predictive performance (RMSE = 0.32, MAE = 0.24) and explained 81% of the variance in academic burnout. Cognitive flexibility emerged as the most influential predictor (38.7% relative importance), followed by school climate (31.2%) and online learning engagement (22.5%). The model exhibited high stability across gender and grade-level subgroups, with explained variance ranging from 78% to 83%.

Conclusion: Academic burnout among Chinese high school students is best explained through a dynamic interaction of cognitive, environmental, and behavioral factors, and the proposed explainable AI framework provides a powerful and transparent tool for early identification and targeted prevention of burnout risk in educational settings.