Where animals avoid inbreeding, different mechanisms of kin discrimination can leave different ‘signatures’ in the patterns of observed mate relationship. For example, consider a species with no paternal care. If a female avoids mating with familiar individuals, one would expect a deficit of offspring whose parents are maternal half-siblings, but paternal half-siblings would be unfamiliar with each other and thus have offspring at the frequency expected by chance. If spatial cues are used to avoid inbreeding, a female would be expected to produce few offspring with males (even unrelated males) living near her birth site. We searched for these and other signatures with data from a long-term study of banner-tailed kangaroo rats, Dipodomys spectabilis, in Arizona, USA, using a combination of intensive censusing, mapping of available dens, microsatellite-based parentage determination, and a randomization routine that determines the numbers of offspring expected if females in the population mate indiscriminately among the males available to them. The data are consistent with the hypothesis that kangaroo rats discriminate kin by familiarity developed via association early in life, rather than by using spatial cues or self-referential phenotype matching. Our approach should be widely applicable as a means of assessing the degree to which kin discrimination exists (in contexts like nepotism as well as inbreeding avoidance) and in inferring what cues animals use to assess categories of relationship.