Males of many species must allocate limited energy budgets between mating and parenting effort. The Challenge Hypothesis provides a framework for understanding these life-history trade-offs via the disparate roles of testosterone (T) in aggression, sexual behavior, and parenting. It predicts that males pursuing mating opportunities have higher T than males pursuing paternal strategies, and in humans, many studies indeed report that men who are fathers and/or pair-bonded have lower T than childless and/or unpaired men. However, the magnitude of these effects, and the influence of methodological variation on effect sizes, have not been quantitatively assessed. We meta-analyzed 114 effects from 66 published and unpublished studies covering four predictions inspired by the Challenge Hypothesis. We confirm that pair-bonded men have lower T than single men, and fathers have lower T than childless men. Furthermore, men more oriented toward pair-bonding or offspring investment had lower T. We discuss the practical meaningfulness of the effect sizes we estimate in relation to known factors (e.g., aging, geographic population) that influence men’s T concentrations.