Background. Self-esteem (SE), a widely used construct in the social sciences, is usually conceptualized as a reflection of socialization and interpersonal experiences that may differ considerably between the genders. Methods. The Rosenberg self-esteem scale was assessed at personal interview in both members of 3793 unselected twin pairs (1517 male–male, 856 female–female and 1420 male–female) from the population-based Virginia Twin Registry. Gender effects on SE were assessed by both analysis of variance and biometrical twin modelling. Results. The mean SE score was slightly but significantly lower in women v. men, and in women who grew up with a male v. a female co-twin. Twin modelling suggested that: (i) individual differences in self-esteem in both men and women were best explained by genetic and individual-specific environment factors; (ii) heritability estimates were similar in women (32%) and in men (29%); and (iii) the same genetic factors that influenced SE in women also influenced SE in men. Analyses supported the validity of the equal environment assumption for SE. The heritability of SE cannot be explained by the moderate correlation between SE and symptoms of depression. Conclusions. These results are inconsistent with prominent gender-related aetiological models for SE, which postulate that individual differences arise from socialization experiences both within and outside the home of origin which differ widely for the two genders. Instead, a significant proportion of the population variance in SE is due to genetically-influenced temperamental variables that are the same in men and women.