Considerable evidence tells us that “being liked” and “being disliked” are related to social competence, but evidence concerning friendships and their developmental significance is relatively weak. The argument is advanced that the developmental implications of these relationships cannot be specified without distinguishing between having friends, the identity of one's friends, and friendship quality. Most commonly, children are differentiated from one another in diagnosis and research only according to whether or not they have friends. The evidence shows that friends provide one another with cognitive and social scaffolding that differs from what nonfriends provide, and having friends supports good outcomes across normative transitions. But predicting developmental outcome also requires knowing about the behavioral characteristics and attitudes of children's friends as well as qualitative features of these relationships.