Strengths and Limitations of a Single-Comparison, Alternate-Response (SCAR) Procedure for Establishing Uni- and Multi-Nodal Stimulus Equivalence Classes

Abstract

Factors determining the yield of establishing stimulus equivalence (SE) classes by means of a single-comparison, alternate-response (SCAR) procedure were explored in a series of experiments. Training involved Pavlovian study phases alternating with response phases until a criterion of performance was attained: there was no trial-by-trial feedback. Factors studied included the explicitness of instructions, types of stimuli used, number of classes to be established, number of nodes, types of tests, exemplar training, and response to an opportunity to relearn. Instructions did not need to be so explicit with 12 or fewer classes as they had been in an earlier study with 48 classes; stimulus type was not critical; but the procedure gave low yields with 3-nodal compared with 1-nodal classes, even with only 3 such classes. Added exemplar training helped to increase yield, but this was still poorer than in an otherwise similar study in which a matching-to-sample (MTS) procedure was substituted for the alternate response requirement. A key distinction may lie in the ambiguity of the alternate response requirement in contrast with the implicit rule in MTS that one of the stimuli present in the comparison array must be correct.