Research on sex differences in antisocial behaviour may shed light on the causes of childhood antisocial behaviour. Using a longitudinal design, we tested whether there were sex differences in the amount of harsh discipline children received or in the effect of harsh discipline and whether this accounted for sex differences in later conduct problems. Our sample was a representative, longitudinal sample of 13,830 twins born in England and Wales between 1994 and 1996. Results showed that boys experienced more harsh discipline than girls and that the sex difference in harsh discipline accounted for 10 percent to 20 percent of the sex difference in conduct problems. We found no evidence that harsh discipline had a greater effect on boys vs. girls. We also found evidence of a bidirectional relationship between harsh discipline and child conduct problems. These findings were replicated within families, automatically controlling for between-families confounding factors. (Contains 3 figures and 4 tables.)