Males can harm the females they interact with, but populations/species vary widely in the occurrence and extent of harm. We consider the merits and limitations of two common approaches to investigating male harm and apply these to an experimental study of divergence in harm. Different physical environments can affect how the sexes interact, causing plastic and/or evolved changes in harm. If harmful male phenotypes are less likely to evolve in situations where females have more control over sexual interactions, populations evolving in environments in which females have greater control should have less harmful males. We test this idea using experimental populations of Drosophila melanogaster that have evolved in either of two environments that vary in the extent to which females can avoid males, or in a third environment without mate competition (i.e., enforced monogamy). We demonstrate an evolved reduction in harm in the absence of mate competition and also in a mate competition environment in which females have greater control. We also show a plastic effect in that otherwise harmful males are no longer so when tested in the environment in which females have greater control. Our results reveal the different perspectives provided by the two methods of studying harm.