The terms `interactivity' and `interactive media' became significant buzz-words during the late 1980s and early 1990s when the multi-media euphoria fascinated politicians, economists, and researchers alike. However, right from the beginning of the scientific debate, the inconsistent usage of the term `interactivity' massively complicated the comparability of numerous empirical studies. This is where this article joins the discussion. First, the article sheds light on the terminological origins of `interactivity' and distinguishes the term from cognate expressions. Further, it restructures and extends existing findings on the basis of a new analysis framework which considers three levels of interactive communication (action level, level of subjective situation evaluation, and level of meaning exchange). Finally, it delivers a systematic overview of specific criteria of interactive communication.