Abstract
Social isolation is an urgent challenge in urban environments worldwide. A widely used strategy is to establish social hubs within local neighbourhoods to promote social connection among residents, especially older adults. Yet little is known about the working and outcome of these social spaces, activities and programmes. The study seeks to address this knowledge gap and examine residents' participation and membership in social activities and organisations within a public housing neighbourhood in Singapore using survey data (N = 300). Developing a comprehensive understanding of the social and spatial factors that facilitate resident interactions, Exponential Random Graph Models are employed, integrating three spatial mechanisms-spatial propinquity, spatial composition, and spatial configuration-alongside residents' sociodemographic profiles, self-reported health and well-being, and perceptions of the local environment (e.g. place attachment, walkability). The study found that integrating multiple spatial mechanisms with residents' characteristics offers deeper insight into local social connectedness. Spatial composition and configuration are stronger predictors of participation and membership than individual traits or proximity. It validates the configurational approach, showing that features like spatial integration and third places shape social interaction patterns. The findings illuminate how specific local organisations function as social hubs, offering guidance for designing community-based interventions while providing tools for evaluating such initiatives.