Learning to be friendly: mediating the tourism–well-being nexus in a small island state

Purpose This study examines the relationship between residents' perceptions of tourism development and their subjective well-being in a small island developing state (SIDS). It explores how a tourism-dominant environment influences well-being through a psychosocial construct, friendliness learning, within the context of New Providence Island, The Bahamas. Design/methodology/approach A quantitative, cross-sectional survey was administered to 374 residents using validated measures of perceptions of tourism development and happiness. Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to evaluate the direct, indirect and mediating effects. Findings The results indicate that perceptions of economic and macro-level tourism development positively influence subjective well-being largely through the mediating role of friendliness learning. In contrast, socio-cultural tourism perceptions were not significant predictors of happiness or friendliness learning. The model explained 26. 8% of the variance in friendliness learning and 20. 9% in happiness, with all significant paths exhibiting positive effects. Research limitations/implications A limitation is that relationships are associations. An implication is that a multi-theoretic framework can explain the psychosocial construct of friendliness learning and its implications for resident happiness in a small island state. Practical implications Practically, fostering local education and prosocial tourism behaviors can enhance residents' happiness while reinforcing sustainable destination development. Originality/value This study introduces friendliness learning as a novel behavioral mechanism linking tourism development to well-being, particularly in archipelagic SIDS contexts. It offers new theoretical insights by integrating psychological empowerment and cultural identity within tourism-well-being models. The structural model showed a good fit and explanatory power (SRMR = 0. 056; R2Friendliness = 0. 268; R2Happiness = 0. 209) with predictive relevance (Q2Friendliness = 0. 233; Q2Happiness = 0. 150).