Abstract:
Recent research concerning choice-making (henceforth referred to as “choice”) analysing behaviour at a local level demonstrated reinforcers primarily signal the subsequent contingency rather than strengthen behaviour. Research about surprise proposed that surprising stimuli facilitate the learning of physically approximate events. This study examined how expected and surprising reinforcers affected choice-making in preschool children. In this experiment, seven children hunted for a preferred item by choosing one of two cups in a series of trials. We manipulated the sequence of reinforcers during baseline so that all reinforcers in a sequence of trials either occurred from choice of the same cup or through alternating choice between cups. We introduced surprising reinforcers after responding stabilized. The results showed that children’s learning of the baseline contingency occurred fast, and control for the next-reinforcer location was strong, although never perfect. Surprising reinforcers had the biggest impact on choice in the immediate trial after. By the second trial, how surprising reinforcers impacted on choice depended on the baseline contingency. Although surprising reinforcers did not have consistent effects, they typically did not cause the participants to repeat the just-reinforced response. Thus, the impact of surprising reinforcers on behaviour in the next trial was not consistent with the strengthening theory. What a surprising reinforcer signalled to behaviour came under the dual control of the most recent reinforcers and more extended reinforcer sequences. There was no evidence suggesting that surprising reinforcers improved learning or discrimination. The delivery of a surprising reinforcer analogizes the mistaken delivery of a reinforcer to behaviour that was not supposed to receive a reinforcer according to the prescribed protocol in a clinical setting. Our results provided evidence that treatment integrity errors compromise the overall learning.