Probability is not an unambiguous concept within the sciences or in vernacular language, yet it is fundamental to much of behavior analysis. The present paper examines some problems this ambiguity creates in general,as well as within the experimental analysis of behavior, in particular. As background material, we first introduce the three most common theories of probability in mathematics and science, discussing their advantages and disadvantages, and their relevance to behavior analysis. Next, we discuss the concept of probability as encountered in (...) the writings of B.F. Skinner and in the contemporary behavior analysis more generally, the latter being based on material drawn from the professional literature and from a questionnaire survey. Although the exercise is basically a descriptive one, we do conclude with some suggestions that may promote more effective action on those occasions when behavior analysts speak of "probability.". (shrink)