Observing Matching-to-Sample Performance and Stimulus Sorting

The primary purpose of the present series of experiments, which includes a total of 155 adult participants, was to determine whether participants observing matching-to-sample (MTS) training and testing only sort stimuli according to experimenter-defined classes. Pooling the results of all four experiments, more than 50% of the participants sorted the stimuli according to experimenter-defined classes with no direct contact with the contingencies in the matching-to-sample training and testing. There was no substantial difference in the outcomes between participants who received a full session of MTS training and testing and those who received a 12-min video clip with training and testing (Experiments 1 and 2). However, the sorting performance was influenced by the ratio of correct responses in the observed MTS training (Experiments 3 and 4). Thus, the condition with the highest ratio of correct responses (80% correct and 20% incorrect) in the MTS training had the most participants sort stimuli according to the experimenter-defined classes. In addition, the participants who sorted correctly also formed equivalence classes in the MTS test (Experiments 2 and 3). Finally, participants who only observed the MTS training (not the test) sorted the stimuli correctly (Experiment 4).