Communication research requires data that are rich enough to contain explicit evidence about processes of communication. The paper fonnally distinguishes among and discusses three kinds of data: aggregational data (such as accumulated in much of psychological inquiries, small group experiments and survey research) and network data (representing observable patterns of interest to relational biology and sociology, for example) are found to be not rich enough to qualify as communication data proper. The paper describes an elementary form of communication data that would contain explicit evidence about the process in question. It exposes some conceptual degenerations in communication research as a consequence of heavy reliance on inadequate data. Pointing to the fact that the advancement of knowledge is critically linked to the ability to process a certain kind of data, the paper concludes that communication research must develop new analytical techniques that are compatible with its most basic concept: communication.