The author reviews and evaluates the principal theoretical measures of attention categorized in three areas: self-report measures, operant behavioral measures, and psychophysiological measures. Self-report measures include a variety of rating scales, interest and attitude scales, Krugman's "number of 'connections,'" and program audience analyzers. Operant behavioral techniques evolve from the conditioning paradigm of Skinner and include response accumulators, tachtiscopic studies, and "shadowing." The psychophysiological measures are connected to attention by the Sokolov theory of the neuronal model. The author discusses these measures in terms of recent studies, methodological pitfalls, and future opportunities for those engaged in communications research and the specific variable of attention. Finally, he develops a mathematical model of attention, based on McPhee's survival theory. The model is constructed on the premise that the brain is a random system and that it is the attentional process that establishes the cognitions from which language and other social behaviors evolve. (Author/RN)