The scientist who uses the experimental form does so in order to explain that which is verified through prediction. The prerequisites for prediction include a universe that is ordered, stable, independent, and knowable. Some assumptions of the predictive argument with respect to communication research are: the operating elements of the message are known, the message is perceived with communality across subjects, the evoking message and the behavior measured or the manipulation used generalize to a class of messages and behaviors or manipulations, the receiver state is known or is equivalent across subjects, the receiver states established during the manipulation can occur in a noncontrolled environment, the behavioral alternatives measured are not dependent solely on the conditions of manipulation, the interaction between the behavior and the entity of record are equivalent across subjects, the entity of record will generalize to a class of such entities or to other classes or recording entities, the known properties of the entity of record are related to the behavior measured, the conditions of manipulation have some equivalence in the noncontrolled environment, and the probability of the noncontrolled environment presenting equivalent conditions is of a significant value. Predictive arguments themselves can be causal, conditional, stochastic, or modeling, and differences among these depend on a priori theory and on statistical analysis. Reports of differences must deal with the importance of the area, the significance of the prediction, and the appropriateness of testing conditions. (TJ)