A set or tendency to anticipate that stimulation will follow a verbal cue suggesting such stimulation is inferred from the magnitude of GSR to the cue. The experimental procedures for modifying the set are analogous to classical conditioning procedures, except that the reinforcing state of affairs is not applied directly to the test cues but is applied instead to a series of seven different cues. Experimental variation is introduced in a 2 x 2 factorial arrangement with one independent variable the number of reinforcements applied before test, and the second independent variable the nature of the test cue. Half of the 80 Ss were tested on a cue involving a feasible suggestion of stimulation. The other half were tested on a cue designated unfeasible. Plots of magnitude of response to the critical test cues as a function of position in the reinforcement series showed an increase of response to both cues at the short training test point, then a decrease with further training. The principal change occurred for the feasible stimulus during the short series of reinforcements.