Opinion leadership is one of the most prominent concepts in communication sciences. A popular method for studying opinion leadership is the use of self-designating scales. The present study examines whether opinion leaders who are identified by self-designating scales literally lead opinions, that is, actually influence others. For this purpose, 78 dyads of friends were interviewed, and their political conversations were observed. The results reveal that self-perceived opinion leadership is associated with aspects of the influence process (i.e., goal to influence, influential communication, and opinion effects) and with other-perceptions of opinion leadership in political conversations. However, given the small effects and the significant role of the interaction partner with regard to the influence process, it is concluded that the extent to which self-designating scales are able to represent the complexity of social reality is limited.