Stimulus equalization: temporary reduction of stimulus complexity to facilitate discrimination learning

Learners with limited behavioral repertoires often have difficulty discriminating complex, multidimensional stimuli. Procedures that use gradual stimulus change have been developed to facilitate such discrimination, but these procedures are often difficult to implement and costly in terms of teacher time and expertise. This study investigated the effectiveness of stimulus equalization, an error reduction procedure involving an abrupt but temporary reduction of dimensional complexity. A microcomputer was used for stimulus presentation, data collection, and response analyses. Preschool children's responding to groups of four stimuli differing along several dimensions was analyzed on four discrimination tasks under several conditions. On each task, one element from a different dimension was programmed as correct. When trial-and-error training failed to establish the discrimination, equalization training began in which differences in the irrelevant dimensions were eliminated. When correct responding developed, the differences were reinstated, and correct performance was maintained in all but one instance. In a repeated acquisition design, stimulus equalization was found to be generally effective and superior to the trial-and-error method. Implications for computer-assisted instruction are discussed.