A influência da preferência por brindes no autorrelato de desempenho em tarefas de leitura

Abstract

This study aimed at investigating the correspondence in a do-say activity. Two experiments were carried out: the first intended to clarify the effects of choice and preference for items on the self-report of children in a reading task performed on the computer; the second experiment endeavored to search for effects of previous reinforcement history using the same task as the former. Six children aged between 7 and 10 took part in Experiment 1. Doing consisted of reading a word out loud and saying comprised the report of having emitted a correct answer or having made a mistake. The entire task was performed on the computer. The item preference was determined by an interview done with the subjects themselves and it was also assessed by a multiple-stimuli without replacement test. The self-report of mistakes and correct answers for each subject was registered during three baseline sessions and no differential consequences followed any particular self-report. Afterwards, the subjects were exposed to the testing sessions with the objects evaluated as of high, medium and low preference, with and without the possibility of choosing what item they wanted to obtain. The total number of sessions was 18 and each of them was planned with 50% of words the subjects did not know how to read. The children were informed that they could only receive the treat (one of the preference items) if they had at least 15 points as the final score what would sum up to 75% or more of right answers for each session. Therefore, to every report of right answer, a point was added to the total score and it was presented on a screen shown in the intertrial interval. The percentage of noncorrespondent reports increased during the testing sessions, if compared to the baseline. It can be noticed a progressive raise on the non-correspondent reports of correct answers for all the subjects, nevertheless, the data were variable. The Experiment 2 aimed to broaden the results of the first and also to evaluate the role of historical variables represented here as a correspondence training carried out before the testing sessions. Three children between 7 and 10 years old were the subjects. The results obtained in the second experiment did not substantially differ from the first. Similar to Experiment 1, the data obtained from Experiment 2 for all children show an increase on emission of self-reports of correct answers, regardless if they made a mistake or not. There were not observable choice effects in both experiments. Moreover, the pattern of responses of most of subjects differed from what was to be expected if the reports were in function of preference. In conclusion, it is of fundamental importance to investigate the role of the score shown during the task as it may have functioned as an immediate conditioned reinforcer during the sessions increasing the emission of reports of correct answers. Keywords: verbal/nonverbal correspondence, reading, preference for consequences, reinforcement.