The article by Smith, Henriques and Parsonson (1985) published in Behaviour Change raises some interesting points, not the least of which is the validity of one of the major conclusions drawn by the authors. They claim “The results of this study indicate that the two treatment conditions employed were not equally effective in increasing the subject's range of supination” (p.56). Putting the clinical utility of their study aside, data available suggest this conclusion is unwarranted.

Essentially, the three conditions (Tml, Tm2, Bl) may be considered to be components of a multiple schedule. Thus, the functional relationship between behaviour and a specific intervention is established if the behaviour co-varies in a predictable fashion with changes in experimenter operations (treatment). On the face of it, the data of Smith et al., appear to meet this criterion.

However, when these same data are plotted in order of presentation, with the data for each session within each condition averaged so as to reduce variability, a different possibility emerges. (See Figure 1.) There is steady improvement in the subject's performance over time, regardless of which treatment condition is in effect in a particular session. These are parallelled with slow increments in baseline response rates. This strongly suggests generalization between conditions or a treatment process common to both interventions.