Moral philosophers draw an important distinction between two kinds of moral responsibility. An agent can be directly morally responsible, or they can be derivatively morally responsible. Direct moral responsibility, so many believe, presupposes that the agent could have behaved differently. However, in some situations, we hold agents responsible even though they could not have behaved differently, such as when they recklessly cause an accident or do not take adequate precautions to avoid harmful consequences. Moral philosophers typically argue that what we ascribe in these cases is derivative moral responsibility. In this paper, I apply this conceptual distinction to the experimental debate about so-called folk-compatibilism. I argue that experimental philosophers have failed to consider this distinction when designing experiments and interpreting their results. I demonstrate that while compatibilism requires judgments of direct moral responsibility, participants in some of the most influential studies ascribed derivative moral responsibility. For this reason, these studies do not speak in favour of compatibilism at all.