Talk-Aloud Protocols during Conditional Discrimination Training and Equivalence Class Formation

The primary focus of the present experiment was to study participants' self-talk during conditional discrimination training and test for the emergence of conditional relations consistent with equivalence. Eighteen adult participants were exposed to either a Many-to-One (MTO) or a One-to-Many (OTM) training structure arranged as a 6-s delayed matching-to-sample (DMTS) procedure. During training and testing, participants were asked to talk aloud. Each trial was divided into three time-windows for recording vocal responses. The first time-window was when the sample was presented, the second time-window was the delay following the offset of the sample stimulus until the onset of the comparisons, and the final time-window was when the comparison array was presented. Overall, participants in the OTM condition talked more than participants in the MTO condition. The participants exposed to MTO increased talking during the third time-window from last part of training to test more than the participants in the OTM condition. For 16 of the 18 participants, talking aloud was less than 55% in the second time-window (during the delay). Conversely, talking in the first and third time-windows were higher than 90% in both groups. Thus, it seems that talking is affected by the shift from training to testing to a greater degree when exposed to MTO than OTM. Furthermore, the data indicate that the participants did not need to talk aloud during the delay to respond correctly on DMTS trials.