The Psychological Record, 2004, 54, 579-602 A COMPARISON OF MATCH-TO-SAMPLE AND RESPONDENT-TYPE TRAINING OF EQUIVALENCE CLASSES MICHAEL C. CLAYTON Jacksonville State University LINDA J. HAYES University of Nevada Throughout the 25-year history of research on stimulus equivalence, one feature of the training procedure has remained constant, namely, the requirement of operant responding during the training procedures. The present investigation compared the traditional match-to-sample (MTS) training with a more recent respondent-type (ReT) procedure. Another consistent feature of the equivalence paradigm is the apparent stipulation that both training and testing must occur before equivalence emerges. In this respect, a more idiosyncratic measure of class acquisition would be desirable. Multidimensional scaling, as a class of exploratory techniques, is introduced as a possible addition to the stimulus equivalence paradigm. Results from 35 subjects in 3 experiments suggest that while the respondent-type training method can be an effective procedure, the operant-based match-to-sample method was clearly more effective in tests for symmetry, equivalence, and extended equivalence. The addition of a scaling procedure proved valuable and showed that both training methods facilitated the emergence of derived relations to varying degrees. Results are evaluated in relation to the importance of broadening the necessary and sufficient training conditions and response requirements for the emergence of stimulus equivalence. When verbally able humans are taught a series of interrelated conditional discriminations, the stimuli involved often become related to each other in untrained ways. For example, when a subject is taught to select a stimulus B in the presence of a stimulus A, and additionally, to select a stimulus C in the presence of stimulus A, it is likely that the subject will also select A in the presence of B (symmetry), A in the presence of C (symmetry), and B in the presence of C and C in the presence of B (equivalence) without explicit training to do so. Portions of this research were conducted as part of Michael Clayton's doctoral research program under supervision of Linda J. Hayes. Much appreciation is owed to Jiang Qian and Dermot Barnes-Holmes for their contributions to this project. We also thank William Dube and an anonymous reviewer for their thoughtful comments. Address correspondence to Michael Clayton, PhD, Psychology Department, Jacksonville State University, 700 Pelham Road North, Jacksonville, AL 36265. (E-mail: mclayton@jsucc.jsu.edu).