The drivers of animal social structures remain poorly understood, particularly in species such as cetaceans that are wide-ranging and challenging to study. Understanding the factors shaping sociality can shed light on population ecology, gene flow, and information transmission. Here, we investigated variation in social structure among three independent island-associated stocks of common bottlenose dolphins (Tursiops truncatus) around the main Hawaiian Islands. We generated social networks for each stock using photo-identification data from 2002 to 2022. We calculated modularity, density, degree centralization, and betweenness centralization to assess network structure. We measured the stocks' available habitat and calculated their population densities. We also quantified association strength with the half-weight association index (HWI) and compared it within- and between-clusters, and by sex for each stock. HWIs revealed that within-cluster associations were much stronger than between-cluster in all stocks. Network modularity and HWI showed the lowest fragmentation into distinct clusters and the strongest associations in the smallest of the three habitats (Kaua‘i-Ni‘ihau). We found no conclusive evidence of sex differences in HWI. Our findings suggest that denser populations might drive social network fragmentation. Our study highlights the importance of further investigating the drivers of sociality.