Sex differences in selection are ubiquitous in sexually reproducing organisms. When the genetic basis of traits is shared between the sexes, such sexually antagonistic selection (SAS) creates a potential constraint on adaptive evolution. Theory and laboratory experiments suggest that environmental variation and the degree of local adaptation may all affect the frequency and intensity of SAS. Here, we capitalize on a large database of over 700 spatially or temporally replicated estimates of sex-specific phenotypic selection from wild populations, combined with data on microclimates and geographical range information. We performed a meta-analysis to test three predictions from SAS theory, that selection becomes more concordant between males and females: (1) in more stressful environments, (2) in more variable environments and (3) closer to the edge of the species' range. We find partial empirical support for all three predictions. Within-study analyses indicate SAS decreases in extreme environments, as indicated by a relationship with maximum temperature, minimum precipitation and evaporative potential (PET). Across studies, we found that the average level of SAS at high latitudes was lower, where environmental conditions are typically less stable. Finally, we found evidence for reduced SAS in populations that are far from the centre of their geographical range. However, and notably, we also found some evidence of reduced average strength of selection in these populations, which is in contrast to predictions from classical theoretical models on range limit evolution. Our results suggest that environmental lability and species range position predictably influence sex-specific selection and sexual antagonism in the wild.This article is part of the theme issue 'Linking local adaptation with the evolution of sex differences'.