The queen-worker conflict over sex allocation provides an excellent opportunity to evaluate the impact of kin selection in social Hymenoptera. Theory predicts that under worker control, colonies with relatively high or low relatedness asymmetry should specialize in producing females or males, respectively. The last qualitative review of this prediction revealed that split-sex ratio was correlated with variation in relatedness asymmetry in 19 out of 25 species (Bourke A. F. G. 2005. Insect evolutionary Ecology. Ed. MDE Fellowes et al. Wallingford: CABI). Here, we perform a meta-analysis to quantify how much of the variance in sex allocation is explained by variation in relatedness asymmetry and to evaluate the potential publication bias. We found only 8 studies containing quantitative data on both sex allocation variation and relatedness asymmetry variation. Additionally 17 studies had data on sex allocation and queen number variation, which may affect both relatedness asymmetry and local resource competition. Relatedness asymmetry variation explains 15% (95% CI: 3-33%) of sex allocation variation among colonies. This value is significantly greater than zero (P < 0.001), but still relatively small when compared to other ecological variable affecting sex allocation (e.g. Local resource competition, West S. A. et al. 2005. Evolution. 59:1211-1228). Queen number explains 6% (CI: 2.4–8.9 %) of sex allocation variation among colonies, which is not statistically different from the previous factor (P = 0.226), but is more difficult to interpret as it potentially combines the effect of relatedness asymmetry and local resource competition. Despite the small number of studies, there was no sign of a publication bias, with no lack of non-significant studies with small sample sizes. This quantitative analysis supports the hypothesis that workers often control sex allocation but also reveals that a large part of the variance remains unexplained.