Abstract Recent studies of non-random paternity have suggested that sperm selection by females may influence male fertilization success. Here we argue that the problems originally encountered in partitioning variation in non-random mating between male competition and female choice are even more pertinent to interpreting patterns of non-random paternity because of intense sperm competition between males. We describe an experiment with the yellow dung fly, Scatophaga stercoraria, designed to partition variance in the proportion of offspring sired by the second male, P2, between males and females, and to control for sperm competition. Large males were shown to have a higher P2 than small males but P2 was independent of the size of the females first mate. This result might suggest an absolute female preference for large males via sperm selection. However, large males have a higher constant rate of sperm transfer and displacement. After controlling for this effect of sperm competition, large males did not achieve higher paternity than small males. We argue that a knowledge of the mechanism of sperm competition is essential so that male effects can be controlled before conclusions are made regarding the influence of sperm selection by females in generating non-random paternity.