Resistance To Change Of Responding To Stimulus Relations

Three experiments assessed resistance to change of response accuracy under different reinforcement rates. Fifteen college students responded to a matching-to-sample task, with one set of baseline conditional relations reinforced at a rich rate and another set reinforced at a lean rate. In two experiments, resistance to change was assessed by presenting tests for emergent equivalence relations and then reversing some of the baseline conditional discriminations and examining changes in the equivalence classes after the reversal; the experiments differed in the presence or absence of overtraining for the rich condition. The third experiment assessed resistance to change by testing for emergent equivalence relations and then requiring participants to respond to distracting stimuli during the matching-to-sample task. Retention after two weeks was measured in the three experiments. The rich reinforcement rate resulted in faster emergence of equivalence responding and greater resistance to distraction only when accompanied by overtraining. The rich reinforcement rate resulted in responding that was more resistant to the reversal of baseline relations for participants who had partial class reversals, but not for participants who had complete class reversals. Furthermore, partial reversals occurred more often when overtraining was minimized or eliminated. There were no systematic differences in retention for the rich and lean conditions in any of the experiments. Findings are discussed in terms of the degree of integration of equivalence classes and the use of conditioned reinforcement to study behavioral momentum in humans.