Using a long-term study of black-tailed prairie dogs (Cynomys ludovicianus), we asked whether subdivision of a subpopulation (colony) into social breeding groups (coteries) influenced gene dynamics. We measured gene dynamics with common statistical tools, F-statistics and effective population size (N_e), but at a finer scale to account for coteries. We used 2 methods of estimating the gene dynamics of subgroups, and determined if these methods produced similar results that were congruent with an empirical measure of the observed effective population size (N_eO). Modified F-statistics were estimated from pre- and postdispersal data from pedigrees and allozymes. Both indicated significant genetic substructuring of the colony subpopulation into coterie breeding groups. The rate of inbreeding of individuals relative to the coterie lineage indicated lower than expected inbreeding at the coterie level. Inbreeding of individuals relative to the colony was consistent with random mating. Asymptotic effective size estimates varied substantially. Chesser's method produced estimates of 77 (range 69–90, pedigree) and 86 (range 70–111, allozyme) individuals consistent with the N_eO of 76 and previous empirical estimates of the instantaneous asymptotic effective size from pedigrees (92.9). Nunney's method produced much lower estimates of approximately one-half the N_eO. Social subdivisions of the colony into coteries clearly influenced gene dynamics. Only the Chesser method accounted for genetic structure introduced by genealogy, both from polygynous mating and matrilines of philopatric females. This may prove important when estimating the rate of loss of genetic variation in highly social mammals.