PremiseNumerous studies have found a positive association between dioecy and polyploidy; however, this association presents a theoretical conflict: While polyploids are predicted to benefit from self‐reproduction for successful establishment, dioecious species cannot self‐reproduce. We propose a theoretical framework to resolve this apparent conflict. We hypothesize that the inability of dioecious species to self‐reproduce hinders their establishment as polyploids. We therefore expect that genera with many dioecious species have fewer polyploids, leading to a negative association between polyploidy and dioecy across genera.MethodsWe used three publicly available databases to determine ploidy and sexual systems for 131 genera and 546 species. We quantified (1) the relationship between the frequency of polyploid species and the frequency of dioecious species across genera, and (2) the proportion of polyploids with hermaphroditism and dioecy across species, adjusting for phylogenetic history.ResultsAcross genera, we found a negative relationship between the proportion of polyploids and the proportion of dioecious species, a consistent trend across clades. Across all species, we found that sexual system (dioecious or not) was not associated with polyploidy.ConclusionsPolyploids are rare in genera in which the majority of species are dioecious, consistent with the theory that self‐reproduction favors polyploid establishment. The low frequency of polyploidy among dioecious species indicates the association is not as widespread as previously suggested. Our findings are consistent with previous studies identifying a positive relationship between the two traits, but only if polyploidy promotes a transition to dioecy, and not the reverse.