Journal of Adolescent and Youth Psychological Studies
Volume:7 Issue: 1, Jan 2026

This study aimed to examine the effectiveness of life skills training in reducing self-handicapping behaviors and their components among master’s students.

 This applied, quantitative study used a semi-experimental pretest–posttest design with a control group and a two-month follow-up. The statistical population comprised all master’s students in Economic Management at the Islamic Azad University, Shiraz Branch, during the 2024–2025 academic year. Using convenience sampling, 78 students were selected and randomly assigned to an experimental group (n = 39) and a control group (n = 39). The experimental group participated in ten 45-minute sessions of life skills training based on Klinke’s (1994) protocol, covering self-awareness, empathy, communication, anger management, problem-solving, stress management, decision-making, creative and critical thinking, and emotional regulation. Data were collected using the Self-Handicapping Questionnaire developed by Jones and Rhodewalt (1978) and analyzed through analysis of covariance (ANCOVA) and multivariate analysis of covariance (MANCOVA) using SPSS 26.

The results revealed significant reductions in total self-handicapping and its components among the experimental group compared to the control group at the posttest stage (p < .001). ANCOVA results indicated strong intervention effects on total self-handicapping (F = 92.34, η² = .591), effort (F = 32.25, η² = .342), and negative mood (F = 21.65, η² = .259), with a moderate effect on excuse-making (F = 8.90, p = .004, η² = .126). The follow-up analysis confirmed the persistence of effects after two months, except for the excuse-making component (p > .05).

Life skills training effectively reduces self-handicapping and its cognitive-emotional components among university students and maintains its impact over time, highlighting its importance as a preventive and developmental intervention in higher education.

The present study was conducted with the aim of developing a cognitive–physical dual-task intervention package and examining its effectiveness on the evidence rate (evidence accumulation speed) in adolescent girls.

 This research employed a quasi-experimental design with pretest–posttest–follow-up and a control group. The statistical population consisted of low-active adolescent girls in Tehran in the 2024–25 academic year. Using convenience sampling, 50 participants were selected and randomly assigned to experimental and control groups (25 in each group). The experimental group received training based on the designed cognitive–physical dual-task package over eight weeks in twenty-four 20-minute sessions. The control group performed moderate-intensity aerobic exercises, starting from below-threshold intensity and progressing to moderately high intensity. To assess cognitive changes, the Wisconsin Card Sorting Test (WCST) was used, with “perseverative errors” and “total errors” serving as indices of cognitive flexibility and evidence accumulation speed. The process of developing the intervention package included designing training lesson plans, a teenager’s workbook, and evaluating content validity using CVR and CVI indices, which ranged from 0.8 to 1 and 0.9 to 1, respectively, confirming the satisfactory validity of the package. Data were analyzed using repeated measures ANOVA.

The results showed that the mean perseverative errors and total errors in the experimental group decreased significantly compared to the control group at posttest and follow-up (p < 0.01). Effect size and eta squared also indicated a considerable impact of the cognitive–physical exercises on increasing the evidence accumulation rate.

The designed package significantly improved cognitive flexibility and decision-making processing speed in adolescent girls. These findings support the effectiveness of integrative mind–body approaches in enhancing fine-grained cognitive indices during adolescence.

This study aimed to compare the paradigmatic models of girls and boys in the formation process of marriage trauma using a grounded theory framework.

This qualitative study employed the systematic grounded theory approach of Strauss and Corbin. Semi-structured, in-depth, face-to-face interviews were conducted with 21 unmarried girls and 17 unmarried boys residing in Shahreza who did not have psychological disorders. Participants were recruited through an Instagram call posted by a psychological clinic and selected via snowball sampling. Interviews lasted 60–80 minutes and explored causal conditions, contextual conditions, intervening conditions, strategies, and consequences related to fear of marriage. Data were analyzed through open, axial, and selective coding to identify core categories and reconstruct gender-specific paradigmatic models explaining marriage trauma.

The analysis revealed distinct gendered pathways in the formation of marriage trauma. For girls, the core category was “the ambiguity of the value of marriage and perfectionism of self and parents,” arising from emotional failures of others, familial expectations, religious–moral incongruence, and economic insecurity. For boys, the core category was “the meaninglessness of marital authenticity,” driven by fear of unsuccessful marriage, economic instability, perceived legal vulnerability, past relational hurt, and sociocultural constraints. Across both groups, intervening conditions such as parental conflict, emotional dependency, and intergenerational narratives intensified trauma trajectories. Strategies included seeking emotional reassurance, emphasizing compatibility, and attempting to learn communication skills; however, consequences such as increased dependency, avoidance, and loss of perceived marriage opportunities were evident.

Marriage trauma among youth arises from intertwined emotional, familial, sociocultural, and structural factors, with clear gender-specific mechanisms shaping fear and avoidance of marriage; understanding these paradigmatic models can inform culturally responsive interventions and support systems.

This study aimed to predict risky behaviors in adolescents based on emotional schemas and rumination.

 This applied, descriptive–correlational study was conducted among 150 male high school students in District 1 of Baharestan during the 2024–2025 academic year. Participants were selected through purposive sampling based on the population structure of the region. Data were collected using three validated instruments: the Risky Behaviors Questionnaire (Zadehmohammadi & Ahmadabadi, 2008), the Emotional Schema Scale—Second Version (Leahy, 2012), and the Rumination Response Scale (Nolen-Hoeksema & Morrow, 1991). After verifying statistical assumptions, inferential analyses were performed using Pearson correlation and simultaneous multiple regression in SPSS-27 to examine the predictive power of emotional schemas and rumination for risky behaviors.

Pearson correlation results indicated significant positive relationships between risky behaviors and rumination, emotional rumination, guilt and shame, and blame, while comprehensibility showed a weak negative correlation. The simultaneous regression model was significant, F(6,143) = 27.99, p = 0.004, demonstrating that emotional schemas and rumination collectively explained 38.8% of the variance in risky behaviors (R = 0.623, R² = 0.388). Among predictors, rumination (β = 0.14, p = 0.01), guilt and shame (β = 0.08, p = 0.04), emotional rumination (β = 0.09, p = 0.05), and blame (β = 0.12, p = 0.01) had significant positive effects, while comprehensibility did not significantly predict risky behaviors (p = 0.11).

The findings highlight the substantial role of emotional schemas and rumination in shaping adolescents’ risky behavioral tendencies, demonstrating that maladaptive emotional beliefs and repetitive negative thinking significantly contribute to engagement in harmful behaviors. Interventions targeting emotional schemas and rumination may effectively reduce risk-taking among adolescents.

 The objective of this study was to determine the effectiveness of assertiveness training on enhancing resilience and empathy among high-school girls in Varamin.

 This quasi-experimental study employed a pre-test–post-test design with a control group. The statistical population included all high-school girls in Varamin, from which 30 participants were selected through convenient sampling. Participants were randomly assigned to experimental and control groups, each consisting of 15 students. The experimental group received assertiveness training across structured educational sessions, while the control group received no intervention during the study period. Data were collected using the Connor–Davidson Resilience Scale (CD-RISC) and the Davis Empathy Questionnaire, both of which demonstrated acceptable reliability in this study (α = 0.86 and α = 0.79, respectively). Descriptive statistics and inferential analyses, including multivariate and univariate analyses of covariance (MANCOVA and ANCOVA), were conducted using SPSS version 25 to compare post-test outcomes while controlling for pre-test scores.

Inferential results showed a significant multivariate effect of group membership on the combined dependent variables of resilience and empathy (Pillai’s Trace = 0.930, F(2, 25) = 164.966, p < 0.001). Between-subject effects indicated that the intervention significantly improved resilience (F = 73.835, p < 0.001) and empathy (F = 232.690, p < 0.001) in the experimental group compared to the control group. Effect sizes were large for both variables (η² = 0.749 for resilience and η² = 0.899 for empathy), and statistical power values were high, demonstrating the robustness of the findings.

The results indicate that assertiveness training is a highly effective intervention for improving both resilience and empathy among high-school girls. Incorporating assertiveness-based programs into school mental health initiatives may support adolescents’ emotional well-being and interpersonal functioning.

The present study was conducted with the aim of determining the effectiveness of this therapeutic approach on feelings of rejection and feelings of inferiority among bereaved adolescent girls.

 This quasi-experimental study employed a pretest–posttest design with a one-month follow-up. The statistical population consisted of bereaved adolescent girls in District 4 of Isfahan. After screening 300 students, 40 individuals with above-average scores were randomly assigned to experimental and control groups (20 participants in each). Data collection instruments included the Perceived Rejection Scale by Penhaligon et al. (2009) and the Inferiority Feeling Questionnaire by Yao et al. (1998). The experimental group received eight 90-minute sessions of Dialectical Behavior Therapy (DBT) based on Linehan’s (1993) protocol, while the control group received no intervention. Data were analyzed using repeated-measures analysis of variance in SPSS-26.

Results of the repeated-measures ANOVA indicated that for both variables—rejection and feelings of inferiority—the main effect of time and the time-by-group interaction were statistically significant (p < .01).

The findings of the present study demonstrate the significant and meaningful effectiveness of Dialectical Behavior Therapy (DBT) in moderating and improving chronic feelings of rejection and feelings of inferiority among bereaved adolescent girls.

 The objective of this study was to predict suicidal ideation in adolescent girls based on resilience, self-compassion, and attachment styles.

 This descriptive–correlational study was conducted among adolescent girls attending public lower-secondary schools in District 16 of Tehran during the 2023–2024 academic year. A total of 239 participants who reported suicidal thoughts in psychological screening were selected through convenience sampling. Data were collected using the Multi-Attitude Suicide Tendency Scale (MAST), the Connor–Davidson Resilience Scale (CD-RISC), the Self-Compassion Scale–Short Form (SCS-SF), and the Revised Adult Attachment Scale (RAAS). Pearson correlation and simultaneous multiple linear regression analyses were performed using SPSS version 26 to examine associations and predictive relationships among the study variables.

Correlation analysis showed that resilience (r = –.36), self-compassion (r = –.14), and secure attachment (r = –.19) were significantly and negatively associated with suicidal ideation, while avoidant (r = .32) and ambivalent/anxious attachment styles (r = .19) exhibited significant positive correlations (p < .05). Regression analysis revealed that the overall model significantly predicted suicidal ideation (F(5,232) = 12.56, p = .001), explaining 21% of the variance (R² = .21). Among the predictors, resilience (β = –.28, p = .001) and avoidant attachment (β = .24, p = .001) were significant predictors, whereas self-compassion, secure attachment, and ambivalent/anxious attachment were not significant predictors.

The findings indicate that resilience and avoidant attachment style play significant roles in predicting suicidal ideation among adolescent girls, underscoring the importance of enhancing adaptive coping resources and addressing maladaptive attachment patterns in suicide prevention efforts targeting this population.

This study aimed to evaluate whether a manualized Moral Reasoning Training (MRT) reduces impulsivity and non suicidal self injury (NSSI) in adolescents more effectively than an active General Life Skills Training (LST) control, and to test whether changes in moral reasoning mediate intervention effects.

 In a single blind, parallel group randomized controlled trial, 240 adolescents (aged 13–17) reporting recent NSSI or elevated impulsivity were randomized to MRT (n = 120) or LST (n = 120). Both interventions comprised eight weekly 90 minute group sessions and were matched for duration and facilitator contact. Primary outcomes were monthly NSSI episode frequency (clinician interview) and trait impulsivity (self report Barratt Impulsiveness Scale); secondary outcomes included principled moral reasoning (DIT 2), behavioral inhibition (Go/No Go), emotion regulation (DERS), depressive symptoms, and global functioning. Assessments occurred at baseline, post treatment (8 weeks), and 3, 6, and 12 months. Analyses followed intent to treat principles using mixed effects models and longitudinal mediation with bootstrapped confidence intervals.

Retention at 12 months was 84%. MRT produced significantly greater reductions in monthly NSSI episodes at post intervention compared with LST (adjusted incidence rate ratio = 0.62, 95% CI 0.50–0.77, p < .001) and larger decreases in trait impulsivity (between group Cohen’s d = 0.42 at post; maintained at 12 months). MRT participants showed larger gains in principled moral reasoning (mean DIT 2 increase ≈ +6.2 vs +1.8) and greater improvement on behavioral inhibition tasks. Longitudinal mediation indicated that DIT 2 change accounted for approximately 45–50% of MRT’s effect on impulsivity and NSSI. Interventions were feasible and safe; no trial related serious adverse events occurred.

Targeted moral reasoning training yielded clinically meaningful and durable reductions in impulsivity and NSSI relative to an active life skills control, with mediation by enhanced principled moral reasoning and corroborating behavioral evidence. MRT represents a promising, mechanism based approach for school delivered prevention of adolescent self injury.

This study aimed to develop and test a causal model of academic buoyancy in gifted secondary school students based on perceived parent–child relationships and teacher–student relationship quality, with the mediating roles of socio-emotional competence and perceived academic pressure.

 The study adopted an applied, descriptive–correlational design using structural equation modeling (SEM). The population consisted of gifted students enrolled in the first and second levels of secondary education in Tehran during the 2023–2024 academic year. A total of 392 students were selected through simple random sampling. Data were collected using standardized questionnaires assessing academic buoyancy, parent–child relationships, teacher–student relationship quality, socio-emotional competence, and perceived academic pressure. Data analysis was conducted using SPSS and AMOS software. Pearson correlation coefficients were calculated, and SEM with maximum likelihood estimation was employed. Indirect effects were tested using bootstrapping with 2,000 resamples to evaluate mediating pathways.

Structural equation modeling indicated that mother–child relationships (β = 0.37, p < .05) and teacher–student relationship quality (β = 0.42, p < .05) had significant direct positive effects on academic buoyancy, whereas the direct effect of father–child relationships was not significant. Socio-emotional competence showed a strong positive direct effect on academic buoyancy (β = 0.45, p < .01), while perceived academic pressure had a significant negative direct effect (β = −0.39, p < .01). Bootstrapping results revealed significant indirect effects of parent–child relationships and teacher–student relationship quality on academic buoyancy through socio-emotional competence (positive) and perceived academic pressure (negative). The overall model demonstrated acceptable goodness-of-fit indices.

The findings highlight academic buoyancy as a relationally embedded and psychologically mediated construct, indicating that supportive parental and teacher relationships enhance gifted students’ vitality primarily by strengthening socio-emotional competence and reducing perceived academic pressure.

The objective of this study was to develop and validate a deep learning model for predicting emotional dysregulation in Colombian adolescents using social media use patterns, sleep quality, and parental attachment styles.

This quantitative, cross-sectional predictive study was conducted among 842 adolescents aged 13–18 years from secondary schools in Bogotá, Medellín, and Cali, Colombia. Participants completed standardized measures of emotional dysregulation, sleep quality, and parental attachment, and objective social media usage metadata were collected via smartphone monitoring over a 14-day period. Data were preprocessed and analyzed using a hybrid convolutional neural network–long short-term memory (CNN–LSTM) architecture. Model performance was evaluated using mean absolute error, root mean square error, coefficient of determination, area under the receiver operating characteristic curve, and classification accuracy. Explainable artificial intelligence techniques were applied to determine the relative importance of predictors.

The deep learning model demonstrated high predictive performance (R² = .79; RMSE = 5.37; AUC = .91; classification accuracy = 87.6%). Emotional dysregulation was significantly associated with social media use (r = .58, p < .001), sleep quality (r = .62, p < .001), and parental attachment (r = −.55, p < .001). Feature importance analysis identified sleep quality (32.4%), nocturnal screen exposure (21.8%), emotional dependency on social media (18.6%), parental attachment security (15.3%), and daily social media duration (11.9%) as the most influential predictors.

The findings indicate that emotional dysregulation in adolescence is strongly shaped by the interaction of digital behavior, sleep processes, and attachment relationships, and that deep learning models offer a powerful tool for early identification of adolescents at elevated emotional risk.

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.

 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.

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%.

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.

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.

 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.

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.

Adolescent depression is shaped by complex, interactive psychosocial and digital factors that are effectively captured through neural network modeling. Strengthening family communication and promoting healthy digital engagement may substantially reduce depressive risk among adolescents.

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.

 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.

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.

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.

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.

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.

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.

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.

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.

 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.

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.

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.

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.

 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.

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.

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.

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.

 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.

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.

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.

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.

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.

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.

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.