Paired Comparison Technique for measuring Stimulus Control Functions in the Pigeon after Training with Discrete Trials

Abstract

SINCE the work of Skinner and his associates¹, free responding techniques, in which an organism can respond repetitively without altering the stimulus situation, have commonly been used in studying discrimination learning in the pigeon. Guttman and Kalish² developed a procedure for obtaining reliable measures of response strength at different “values” along a “stimulus dimension” (for example, wavelength of a patch of light); from these, “stimulus control functions” could be plotted. This procedure has been used by Hanson³ to obtain post-discrimination stimulus control functions, and later workers have used a wide variety of experimental designs and stimulus dimensions⁴. In a free responding situation however, as Jenkins has pointed out⁵, the assessment of stimulus control may be subject to error from confounding effects—for example, a reinforcement within a stimulus presentation may in itself lead to an increased rate of responding on that presentation. He suggests that these effects can be overcome most easily by the use of a discrete trial training procedure, in which the organism's response terminates the trial. But the free responding Guttman and Kalish procedure for measuring stimulus control functions then becomes inappropriate.