Journal of Adolescent and Youth Psychological Studies (JAYPS)

Objective: The present study aimed to develop and test a predictive model examining the extent to which cognitive flexibility explains individual differences in creative problem-solving among adolescents beyond executive functioning and demographic variables.

Methods and Materials: This quantitative cross-sectional study was conducted with a sample of 438 adolescents aged 13–17 years recruited from secondary schools in Spain using multistage cluster random sampling. Cognitive flexibility was assessed using the Spanish adaptation of the Cognitive Flexibility Inventory, and creative problem-solving was measured through a combined performance-based task battery and self-report scale evaluating fluency, flexibility, originality, and elaboration. Executive functioning variables, including working memory and inhibitory control, were measured using standardized teacher-report instruments, and academic achievement data were obtained from school records. Data analysis included descriptive statistics, Pearson correlations, hierarchical multiple regression, and structural equation modeling. In addition, a machine learning random forest regression model with 10-fold cross-validation was applied to compare predictive accuracy with traditional regression approaches.

Findings: Cognitive flexibility demonstrated a strong positive association with creative problem-solving (p < 0.001). Hierarchical regression analyses revealed that cognitive flexibility significantly increased the explained variance in creative problem-solving beyond demographic and executive control variables, with the final model accounting for 47 percent of the variance (ΔR² = 0.26, p < 0.001). Both flexibility dimensions emerged as significant predictors (p < 0.001). Structural equation modeling confirmed a strong latent path from cognitive flexibility to creative problem-solving (β = 0.68, p < 0.001), with the overall model demonstrating excellent fit indices and explaining 52 percent of the variance. Machine learning analyses corroborated the dominant predictive contribution of cognitive flexibility.

Conclusion: The findings indicate that cognitive flexibility functions as a central and robust predictor of creative problem-solving in adolescence, integrating executive control processes and contributing substantial incremental validity.

Objective: The present study aimed to develop and validate an explainable machine learning model capable of accurately predicting school dropout risk by integrating psychosocial and cognitive variables.

Methods and Materials: A cross-sectional predictive design was employed with a sample of 1.172 secondary school students from three federal states in Germany. Standardized instruments were used to assess psychosocial variables including depressive symptoms, academic self-efficacy, school belonging, teacher and peer support, self-regulation, and academic motivation. Cognitive performance was measured through computerized tasks assessing working memory, processing speed, and fluid reasoning. Socioeconomic status and migration background were included as contextual covariates. Data preprocessing involved multiple imputation, normalization, and class imbalance correction using SMOTE. The dataset was partitioned into training, validation, and independent test subsets. Multiple supervised learning algorithms—logistic regression with elastic net regularization, support vector machines, random forest, and gradient boosting—were trained and compared using cross-validated hyperparameter optimization. Model performance was evaluated using AUC, balanced accuracy, F1-score, and calibration indices. Explainability was ensured through SHAP-based global and local feature attribution analyses.

Findings: Gradient boosting achieved the highest predictive performance (AUC = 0.92; balanced accuracy = 0.86), significantly outperforming linear models. Psychosocial variables demonstrated stronger predictive power than cognitive variables alone, yet the integration of both domains significantly improved overall model accuracy. Depressive symptoms, academic self-efficacy, and school belonging emerged as the most influential predictors, while processing speed and working memory provided incremental predictive validity. Nonlinear threshold effects were observed, indicating that elevated emotional distress and reduced cognitive efficiency substantially increased dropout probability.

Conclusion: The findings demonstrate that explainable machine learning models integrating psychosocial and cognitive indicators can reliably predict dropout risk while preserving interpretability.

Objective: The present study aimed to develop and interpret explainable machine learning models examining the nonlinear and interactive associations between attachment insecurity dimensions and digital dependency among adolescents.

Methods and Materials: This cross-sectional quantitative study was conducted among 742 secondary school adolescents aged 13–18 years recruited through multistage cluster sampling from urban regions. Participants completed standardized self-report measures assessing attachment anxiety and attachment avoidance, digital dependency, and demographic variables including daily screen time. Data were analyzed using Explainable Boosting Machines (EBMs), a transparent generalized additive modeling approach enhanced with gradient boosting. The dataset was divided into training and testing subsets, and model performance was evaluated using R², RMSE, and MAE indices. Five-fold cross-validation was implemented to optimize hyperparameters. Global feature importance scores, partial dependence functions, and interaction analyses were generated to examine nonlinear thresholds and synergistic risk patterns.

Findings: The EBM model explained a substantial proportion of variance in digital dependency scores (R² = 0.47) with satisfactory predictive accuracy. Attachment anxiety emerged as a strong psychological predictor, demonstrating a nonlinear threshold effect in which digital dependency increased sharply at moderate levels of anxiety and plateaued at very high levels. Attachment avoidance showed a weaker but significant positive association with digital dependency, characterized by a gradual linear trend. Daily screen time was the most influential predictor overall, and a significant interaction effect was identified between attachment anxiety and screen time, indicating amplified dependency risk among adolescents high on both variables.

Conclusion: The findings underscore the central role of attachment insecurity—particularly attachment anxiety—in shaping adolescents’ vulnerability to digital dependency and demonstrate the utility of explainable machine learning approaches for identifying nonlinear developmental risk patterns.

Objective: The present study aimed to examine whether machine learning–detected online social exclusion predicts adolescent loneliness and whether computationally derived exclusion indicators provide incremental explanatory power beyond self-reported perceived exclusion.

Methods and Materials: A cross-sectional correlational design was employed with a sample of 718 adolescents aged 13–17 years recruited from secondary schools in central Mexico. Participants completed the Spanish-adapted UCLA Loneliness Scale and a validated self-report measure of online social exclusion. In addition, anonymized digital interaction data from the previous 30 days were collected and processed using natural language processing techniques. A supervised transformer-based machine learning model was trained to classify exclusionary linguistic and interactional patterns within 96,438 message entries. Individual-level online exclusion probability scores were computed based on linguistic markers, response latency asymmetry, and unanswered message ratios. Hierarchical regression analyses and structural equation modeling were conducted to test direct and indirect associations between machine learning–detected exclusion, perceived exclusion, and loneliness.

Findings: Machine learning–detected online social exclusion was positively and significantly associated with adolescent loneliness (p < 0.001). Hierarchical regression analysis demonstrated that computationally detected exclusion predicted loneliness above and beyond demographic variables and self-reported exclusion (ΔR² = 0.06, p < 0.001). Structural equation modeling indicated acceptable model fit and revealed that perceived exclusion partially mediated the relationship between machine learning–detected exclusion and loneliness (indirect effect p < 0.001), while the direct path remained significant (p < 0.001).

Conclusion: The findings indicate that algorithmically detected online exclusionary patterns constitute a significant and independent predictor of adolescent loneliness. Integrating machine learning–based behavioral analytics with psychosocial assessment enhances the precision of loneliness risk identification and offers promising avenues for early detection and prevention strategies in digitally mediated youth environments.

Objective: The present study aimed to predict cyberbullying perpetration among adolescents using Random Forest modeling of moral disengagement mechanisms and empathy deficits.

Methods and Materials: This cross-sectional quantitative study was conducted among 742 secondary school students (ages 13–18 years) from three provinces in South Africa using multi-stage cluster sampling. Participants completed validated self-report instruments measuring cyberbullying perpetration, moral disengagement, and empathy deficits, alongside demographic indicators and daily internet usage. Data were screened, cleaned, and randomly divided into training (70%) and testing (30%) datasets. A Random Forest classifier with 500 trees was trained to distinguish high versus low cyberbullying perpetration. Hyperparameters were optimized using cross-validation. Model performance was evaluated through accuracy, precision, recall, F1-score, and area under the receiver operating characteristic curve (AUC-ROC). A logistic regression model was estimated as a baseline comparator. Variable importance indices and partial dependence plots were generated to examine predictor contributions and non-linear interaction patterns.

Findings: Inferential analyses indicated that moral disengagement and empathy deficits were significant predictors of cyberbullying perpetration (p < .001). Male students reported higher levels of cyberbullying and moral disengagement (p < .001). The Random Forest model outperformed logistic regression, achieving superior classification accuracy (0.86 vs. 0.74) and AUC-ROC (0.91 vs. 0.78). Variable importance metrics identified overall moral disengagement, dehumanization, and attribution of blame as the strongest predictors, followed by empathy deficits. Partial dependence analysis revealed non-linear threshold effects, with sharp increases in predicted cyberbullying probability at higher levels of moral disengagement, particularly when combined with elevated empathy deficits.

Conclusion: The findings demonstrate that cyberbullying perpetration is most strongly predicted by moral disengagement mechanisms and empathy deficits.

Objective: The present study aimed to model the symptom-level network structure of anhedonia and reward responsivity in Moroccan adolescents with subclinical depression.

Methods and Materials: A cross-sectional quantitative design was employed with a community sample of 468 adolescents aged 14–18 years recruited from public high schools in Rabat, Casablanca, and Fez, Morocco. Participants met criteria for subclinical depressive symptoms based on standardized screening thresholds while not fulfilling diagnostic criteria for major depressive disorder. Measures included the Snaith–Hamilton Pleasure Scale (SHAPS) to assess anhedonia, the Temporal Experience of Pleasure Scale (TEPS) to capture anticipatory and consummatory pleasure, and the Behavioral Activation System (BAS) Reward Responsiveness subscale. Data were analyzed using Gaussian Graphical Modeling with graphical LASSO regularization and EBIC model selection. Centrality indices (strength, closeness, betweenness) and bridge centrality metrics were computed. Network accuracy and stability were evaluated using nonparametric bootstrapping procedures and correlation stability coefficients.

Findings: Network estimation revealed a stable and densely interconnected structure among reward-related symptoms. Anticipatory pleasure deficits demonstrated the highest strength centrality, indicating their structural prominence within the network. Significant negative partial correlations were observed between anhedonia and both anticipatory and consummatory pleasure, as well as reward responsiveness. Bridge centrality analyses identified reduced anticipatory excitement and diminished emotional reactivity to positive stimuli as key connectors linking anhedonia and reward responsivity clusters. Bootstrapped confidence intervals supported the robustness of edge weights, and the correlation stability coefficient for strength centrality exceeded recommended thresholds, confirming network reliability.

Conclusion: The findings suggest that anticipatory motivational deficits occupy a central and bridging position within the symptom architecture of subclinical adolescent depression.

Objective: The present study aimed to examine, through structural equation modeling, the direct predictive effect of intolerance of uncertainty on anxiety sensitivity among early adolescents.

Methods and Materials: This quantitative cross-sectional study was conducted with a sample of 428 early adolescents aged 11 to 14 years enrolled in public middle schools in Turkey. Participants were selected using multistage cluster sampling from urban districts representing middle socioeconomic status. Data were collected using standardized self-report measures assessing intolerance of uncertainty and anxiety sensitivity, both of which demonstrated satisfactory internal consistency in the present sample. After preliminary screening for missing data, normality, and outliers, confirmatory factor analyses were conducted to validate the measurement models. Structural equation modeling with maximum likelihood estimation was then employed to test the hypothesized direct path from intolerance of uncertainty to anxiety sensitivity. Model fit was evaluated using multiple goodness-of-fit indices, including chi-square to degrees of freedom ratio, Comparative Fit Index, Tucker–Lewis Index, Root Mean Square Error of Approximation, and Standardized Root Mean Square Residual.

Findings: The measurement models demonstrated acceptable to good fit indices, confirming the adequacy of the latent constructs. The structural model also showed satisfactory fit to the data. Intolerance of uncertainty significantly and positively predicted anxiety sensitivity (β = .63, p < .001). The model explained 40 percent of the variance in anxiety sensitivity, indicating a substantial effect size. All standardized factor loadings were statistically significant and within acceptable ranges, supporting the reliability of the latent structure.

Conclusion: The findings indicate that intolerance of uncertainty functions as a strong cognitive vulnerability factor predicting anxiety sensitivity in early adolescence. These results support transdiagnostic models of anxiety and highlight the importance of targeting intolerance of uncertainty in early preventive and intervention efforts aimed at reducing anxiety-related sensitivity and emotional distress during adolescence.

Objective: The present study aimed to examine longitudinal trajectories of brooding and reflective rumination and to determine their parallel predictive effects on changes in suicidal ideation among high-risk adolescents using a latent growth curve modeling framework.

Methods and Materials: A prospective four-wave longitudinal design was employed with a sample of 362 high-risk adolescents aged 13–17 years recruited from secondary schools and youth mental health services. Participants completed validated self-report measures of brooding and reflective rumination, suicidal ideation, and depressive symptoms at three-month intervals over a 12-month period. Measurement invariance across time was established prior to growth analyses. Unconditional latent growth curve models were first estimated separately for brooding, reflection, and suicidal ideation to assess intercepts and linear slopes. A parallel process latent growth model was subsequently specified to examine whether baseline levels and rates of change in rumination subtypes predicted trajectories of suicidal ideation while controlling for depressive symptoms and gender. Model fit was evaluated using CFI, TLI, RMSEA, and SRMR indices, and missing data were handled using full information maximum likelihood estimation.

Findings: Results indicated significant positive linear growth in brooding and suicidal ideation over time, with substantial inter-individual variability in slopes. Reflective rumination demonstrated a stable mean trajectory but significant slope variance. In the parallel process model, the brooding intercept significantly predicted the suicidal ideation intercept (β = .41, p < .001), and the brooding slope significantly predicted growth in suicidal ideation (β = .38, p < .001). Reflective rumination did not significantly predict changes in suicidal ideation trajectories. Depressive symptoms remained a significant covariate of baseline suicidal ideation. Overall model fit indices indicated good structural fit.

Conclusion: Findings identify brooding, but not reflective rumination, as a dynamic longitudinal risk mechanism contributing to escalating suicidal ideation among high-risk adolescents, underscoring the importance of targeting maladaptive rumination processes in prevention and intervention efforts.