The evolution of female multiple mating is a highly controversial topic, especially in social insects. Here we analyze, using comparative analyses and simulation models, the merits of two major contending hypotheses for the adaptive value of polyandry in this group. The hypotheses maintain that, respectively, the resulting genotypic diversity among offspring within a colony: (1) mitigates against the effects of parasites; or (2) favors adaptive division of labor. Only two of 11 phylogenetically uncontrolled comparative analyses supported an association between polyandry and the complexity of division of labor (measured here using worker caste polymorphism or polyethism) as proposed by hypothesis 2, and after controlling for phylogeny there were no significant associations. In contrast, a previous study demonstrated such an association for parasite load as expected under hypothesis 1. In addition, we used simulation models to track the spread of an initially rare allele for double mating in a population of single-mating alleles, thus analyzing the crucial first step from monandry to polyandry. We find that double mating evolves consistently under antagonistic coevolution given that parasites exert sufficient selection intensity. In contrast, selection for enhanced division of labor resulted in only an erratic appearance of polyandry in highly (and mostly negatively) autocorrelated environments where no coevolutionary dynamics were allowed. Together, we interpret these results to suggest that parasites, and the antagonistic coevolutionary pressures they exert, may play an important role in the evolution of polyandry in social hymenopteran populations.