Maternal stress during gestation has the potential to affect offspring development via changes in maternal physiology, such as increases in circulating levels of glucocorticoid hormones that are typical after exposure to a stressor. While the effects of elevated maternal glucocorticoids on offspring phenotype (i.e., “glucocorticoid-mediated maternal effects”) have been relatively well established in laboratory studies, it remains poorly understood how strong and consistent such effects are in natural populations. Using a meta-analysis of studies of wild mammals, birds, and reptiles, we investigate the evidence for effects of elevated maternal glucocorticoids on offspring phenotype and investigate key moderators that might influence the strength and direction of these effects. In particular, we investigate the potential importance of reproductive mode (viviparity vs. oviparity). We show that glucocorticoid-mediated maternal effects are stronger, and likely more deleterious, in mammals and viviparous squamate reptiles compared with birds, turtles, and oviparous squamates. Effects on offspring traits were more strongly negative when measured at or close to birth; no other moderators (timing and type of manipulation, or type of trait measured) were significant predictors of the strength or direction of the phenotypic effects on offspring. These results provide evidence that the evolution of a prolonged physiological association between embryo and mother sets the stage for maladaptive, or adaptive, prenatal stress effects in vertebrates driven by glucocorticoid elevation.