Increasing awareness that disorders of attention may underlie cognitive dysfunctions associated with intoxication and neurodegenerative disease has stimulated research into the neural bases of attention. Because attention comprises a constellation of hypothetical cognitive processes, it can only be inferred from behavior, of either human or non-human subjects, under appropriate experimental conditions. Many behavioral procedures have been proposed for modeling attention in animals, but not all of these procedures have been systematically associated with specific attentional processes. This review endeavors to evaluate critically the construct validity of these procedures (i.e., to determine the degree to which a given procedure assesses a particular process) and to suggest experiments to improve the conceptual links between these procedures and the processes they purport to assess. Five categories of processes have been identified from the animal literature: orienting, expectancy, stimulus differentiation (including stimulus salience, discrimination of a critical stimulus from its context, and selection among stimuli), sustained attention, and parallel processing. The review discusses the strengths and weaknesses of specific behavioral procedures for assessing these categories of attentional processes and, given the conceptual uncertainties involved, it attempts to summarize the present state of knowledge of the pharmacology and neurobiology of attention.