A machine learning study of the predictors of fear of happiness in Turkey and the USA

Abstract

Objective: To identify key predictors of Fear of Happiness (FOH) by examining 22 psychological and demographic variables in Turkish and American samples using machine learning. Methods: Random forest analyses were conducted on cross-sectional data from Turkish (N = 824) and American (N = 973) participants to estimate the relative importance of personality traits, attachment patterns, emotion regulation difficulties, existential beliefs, and demographic factors in predicting FOH. Results: Existential nihilism and difficulties in emotion regulation emerged as the strongest predictors of FOH in both cultures, followed by insecure attachment styles (anxious and avoidant) and perfectionism. Predictive performance was higher in the American sample. Culture-specific differences were observed, with neuroticism showing greater relative importance in the American sample and loneliness showing greater relative importance in the Turkish sample. Demographic variables (gender, age, and education) and ideological beliefs (religiosity and fatalism) showed minimal importance in both cultures. Conclusions: The findings indicate that psychological variables (particularly existential nihilism, emotion regulation difficulties, and attachment styles) are more strongly associated with FOH than demographic or ideological factors. These results contribute to a clearer understanding of FOH across cultures, provide data-driven insights, and inform the development of hypotheses for future research.