Steep socio-economic gradients in family size were a major source of disparities for children in the early 20th century and prompted much social research and public commentary. By the 1960s, a scholarly consensus was emerging that SES differentials in women’s fertility in western countries were tending to narrow but developments since then have received limited attention and a children’s perspective relating to the distinct question of sibling numbers (or 'sibsize') has been lacking. Drawing mainly on data from the United States but with some comparative information for other western countries, this paper finds that a sharp reduction in social disparities in sibsize occurred in the final third of the twentieth century and acted as an important (though in the US case, incomplete) social leveller for children. This development is significant as a counter to other aspects of sociodemographic change in the same period which have been found to widen social inequalities for children. A key implication is that until we pay closer attention to sibsize patterns, our picture of how socio-demographic change has affected social inequalities among children in recent decades may be both incomplete and unduly negative.