Three experiments were used to investigate individuals' hypothesis-testing process as a function of moral perceived utilities, which in turn depend on perceived responsibility and fear of guilt. Moral perceived utilities are related to individuals' moral standards and specifically to people's attempt to face up to their own responsibilities, and to avoid feeling guilty of irresponsibility. The results showed that responsibility and fear of guilt in testing hypotheses involved a process defined as prudential mode, which entails focusing on and confirming the worst hypothesis, and then reiterating the testing process. In particular, the results showed that responsible and guilt-fearing individuals: (1) tended to search prudentially for examples confirming the worst hypothesis and to search for counter-examples falsifying the positive hypothesis; (2) focused on the worst alternative, and tended to confirm it; (3) prudentially kept up the testing process, even if faced with initial positive evidence. Our discussion of the results emphasises how people are largely pragmatic in their hypothesis testing, using efficient cognitive strategies that focus on error minimisation rather than on truth detection. In a context of responsibility and guilt, the errors are linked to people's failure to face up to their own responsibilities, and are thus moral errors.