Sexual dimorphism in the human pelvis has been studied widely for forensic purposes, but it is still unclear to what extent it varies among human populations. There is evidence that microevolutionary processes, both neutral (i.e., population history) and selective (e.g., thermoregulatory adaptation and size-related obstetrical constraints) contribute to explain pelvic variation among populations, but the extent to which these factors affect pelvic sexual dimorphism is unknown. In this study, I analyze sexual dimorphism of the os coxae in 20 globally distributed human populations, using 3D morphometric data to separate the size and shape components of sexual differences. After evaluating population differences in the degree and pattern of sexual dimorphism, I test for the effect of population history, climate, and body size in shaping global diversity. The results show that size and shape dimorphism follow different patterns. Coxal size dimorphism is generally quite consistent through populations, with males bigger than females, but it appears to be reduced in small-bodied populations, possibly in relation to obstetrically-related selective pressures for a spacious birth canal. Beyond a general species-wide pattern of shape dimorphism, commonly used for forensic sex determination, other aspects of sexual differences in coxal shape vary among human populations, reflecting the effects of neutral demographic processes and climatic adaptation.