Social relationships endow health and fitness benefits, but considerable variation exists in the extent to which individuals form and maintain salutary social relationships. The mental and physical health effects of social bonds are more strongly related to perceived isolation (loneliness) than to objective social network characteristics. We sought to develop an animal model to facilitate the experimental analysis of the development of, and the behavioral and biological consequences of, loneliness. In Study 1, using a population-based sample of older adults, we examined how loneliness was influenced both by social network size and by the extent to which individuals believed that their daily social interactions reflected their own choice. Results revealed three distinct clusters of individuals: (i) individuals with large networks who believed they had high choice were lowest in loneliness, (ii) individuals with small social networks who believed they had low choice were highest in loneliness, and (iii) the remaining two groups were intermediate and equivalent in loneliness. In Study 2, a similar three-group structure was identified in two separate samples of adult male rhesus monkeys (Macaca mulatta) living in large social groups: (i) those high in sociability who had complex social interaction with a broad range of social partners (putatively low in loneliness), (ii) those low in sociability who showed tentative interactions with certain classes of social partners (putatively high in loneliness), and (iii) those low in sociability who interacted overall at low levels with a broad range of social partners (putatively low or intermediate in loneliness). This taxonomy in monkeys was validated in subsequent experimental social probe studies. These results suggest that, in highly social nonhuman primate species, some animals may show a mismatch between social interest and social attainment that could serve as a useful animal model for experimental and mechanistic studies of loneliness.