The effort to understand the genetic basis of human sociality has been encouraged by the diversity and heritability of social traits like cooperation. This task has remained elusive largely because most studies of sociality and genetics use sample sizes that are often unable to detect the small effects that single genes may have on complex social behaviors. The lack of robust findings could also be a consequence of a poor characterization of social phenotypes. Here, we explore the latter possibility by testing whether refining measures of cooperative phenotypes can increase the replication of previously reported associations between genetic variants and cooperation in small samples. Unlike most previous studies of sociality and genetics, we characterize cooperative phenotypes based on strategies rather than actions. Measuring strategies help differentiate between similar actions with different underlaying social motivations while controlling for expectations and learning. In an admixed Latino sample (n = 188), we tested whether cooperative strategies were associated with three genetic variants thought to influence sociality in humans-MAOA-uVNTR, OXTR rs53576, and AVPR1 RS3. We found no association between cooperative strategies and any of the candidate genetic variants. Since we were unable to replicate previous observations our results suggest that refining measurements of cooperative phenotypes as strategies is not enough to overcome the inherent statistical power problem of candidate gene studies.