Group-living behavior is taxonomically widespread but rare in spiders. The conventional view is that the main pathways to group-living in spiders are either sub-social, where extended maternal care leads to prolonged sibling association; or communal living, where individuals aggregate to exploit a common resource. Female egg-sac guarding behavior occurs throughout kleptoparasitic spiders in the subfamily Argyrodinae (Theridiidae), while individuals in group-living species cohabit in the resource rich webs of their host spiders. These attributes fit both sub-social and communal routes to group-living, which offers new insights to study the early stages of social evolution. We investigated whether members of kleptoparasitic groups in natural populations comprise related individuals by comparing the population structure of two group-living species, Argyrodes miniaceus and A. cf. fissifrons, and two solitary species, A. fasciatus and Neospintharus trigonum. We found that: (1) genetic-spatial autocorrelation in group-living species was highest among spiders sharing the same host web and declined steeply with increasing distance, but no significant autocorrelation at any scale for solitary species; (2) there was high relatedness among group members in two cases of group-living species, which indicated relatedness was not an adhesive agent in most of the groups, but no high relatedness in solitary species; and (3) the host web boundary was not the sole predictor of genetic structures in group-living species. These results suggest that population genetic structure in the group-living species is caused by limited dispersal of group members that is favored by ecological conditions, including the nature and size of resources. In contrast, the absence of genetic structuring in populations of solitary species indicates a high level of dispersal with individual interactions unlikely to have fitness benefits.