Akademik Veri Yönetim Sistemi
Aging and Mental Health, 2026 (SCI-Expanded, SSCI, Scopus)
Objectives: This study examines the multidimensional determinants of perceived ageism using a combined theory- and data-driven framework based on nationally representative Türkiye Elderly Profile Survey (2023) data, identifying age-specific risk and protective factors through an index-based measure. Method: Using a cross-sectional design, perceived ageism was measured with an 11-item scale split into two sub-dimensions via exploratory factor analysis. Indicators were normalized into composite indices assessed by Cronbach’s alpha and tetrachoric EFA. The analysis combined TD OLS models across five age groups with DD machine learning (XGBoost and gradient boosting). Performance was evaluated using precision, recall, F1-score, and ROC-AUC, with SHAP improving interpretability. Results: Geriatric depression was the strongest risk factor across all age groups, both ageism dimensions, and both analytical methods. Social participation showed the strongest protective effect against interpersonal ageism (β = −0.154; −0.232 in the 75+ group). Digital competence protected both dimensions in the 65–74 group but lost its interpersonal effect in the 75+ group (β = −0.028, p = 0.366). The two-dimensional approach revealed suppression effects unseen in aggregate models: education had opposite effects across dimensions, and social participation differed markedly between structural and interpersonal ageism. Conclusion: Perceived ageism reflects psychosocial vulnerability, social exclusion, digital disadvantage, and healthcare inequities, underscoring the value of integrated analytical frameworks for age-sensitive policy design.