Demographics, health, socioeconomic status, values, beliefs, and attitudes are important predictors of subjective well-being. However, the relative importance of these multi-domain predictors remains unclear. Drawing on two large-scale cross-national adult samples ( N ₁ = 148,105, M age = 45.528, SD = 17.165; N ₂ = 147,644, M age = 45.510, SD = 17.163), this study employed machine learning to evaluate the relative contributions of predictors across these domains in predicting life satisfaction and happiness. Results identified both shared and outcome-specific predictive patterns. Locus of control and state of health emerged as the most important predictors, with locus of control ranking highest for life satisfaction and state of health for happiness. Additional key predictors included household income, marital status, in-group trust, acceptance of homosexuality, and age, although their relative importance differed across outcomes. Cross-cultural analyses showed that socioeconomic status contributed more to subjective well-being in collectivistic cultures, whereas marital status was more important in individualistic cultures. Typical profiles indicate that individuals with stronger internal locus of control, better health, higher household income, greater in-group trust, being partnered, and more accepting attitudes toward homosexuality are more likely to be satisfied with their lives and happy. Older individuals are more likely to be satisfied with their lives, whereas the association between age and happiness is complex and potentially nonlinear. By systematically ranking predictors across multiple domains, this study provides an integrated perspective on the relative contributions of objective life circumstances and psychosocial orientations, offering guidance for interventions and policies aimed at promoting subjective well-being.
Li et al. (Thu,) studied this question.