Functional analysis and behavioral variability: some considerations

Abstract

This paper analyses the concept of variability and the role it has been playing in the experimental analyses of behavior. A formal definition of functional analyses is initially presented. Taking into account Sidman’s considerations, a first variability definition is given. It is demonstrated that variability violates the definition of functional analysis and becomes a problem to the empirical investigation of behavior. It is claimed, however, that behavioral variability can become, itself, subject matter of behavioral research. It is demonstrated that concept of variability changes, when variability becomes a dependent variable. Some empirical works are described which took this approach. From Sidman´s point of view, to investigate behavioral variability implies to investigate the ciclic properties of behavior. It is argued, however, that behavioral studies can be wider than that. The concept of variability is retaken and a more general definition of the term is proposed. It is claimed that the investigation of behavioral variability is a legitimate one, even though it is refused that variability is a fundamental dimension of behavior. We conclude that there have been three ways to deal with the variability of behavior: 1-it is attempted to eliminate variability, by identifying the responsible variables; 2- It is possible also to investigate the variability itself, by introducing it into the formulation of behavioral laws; 3- When the variability is not a problem of a experimental technique nor a datum, it is incorporated into the knowledge through the concept of response class.