This paper explores how starting school at a younger age affects the developmental score gaps between relatively advantaged and disadvantaged children. While previous findings suggest that delaying school entry may improve school readiness, less is known about whether it has differential effects for advantaged and disadvantaged children. For disadvantaged children, starting school early may be a better alternative to staying at home for longer as school provides a more stable and educational environment than the family home, overcompensating for the penalties of starting school early. This may be less applicable to relatively advantaged children who generally have greater access to resources in the home and who are more likely to utilise formal pre-school services. We use the Longitudinal Study of Australian Children to investigate if there is support for this hypothesis. The endogeneity of school starting age is addressed using the regression discontinuity design. We find that an early school start generally improves children s cognitive skills, which is even more pronounced for disadvantaged children. In contrast, an early school start tends to negatively affect children s non-cognitive skills with both advantaged and disadvantaged children affected in similar ways. Thus, our findings suggest that an earlier school entry may narrow the gaps in cognitive skills, whereas the gaps in non-cognitive skills are not affected by the school starting age.