Auto-relato de erros em tarefas de leitura: efeitos de um treino de correspondência

Abstract

When children with a history of school failure report on the outcomes of reading responses, most of them tend to report mostly correct responses, even after they make mistakes. This study investigated variables influencing correspondence in these reports and attempted to train correspondence to ensure accurate reports of errors, as well as of correct responses. Experimental sessions presented series of words on a computer screen. The computer dictated the correct word and children selected a green or a red window to report that the response had been correct or wrong. Baseline sessions showed that reports of errors as correct responses increased as a function of error probability. Training sessions then reinforced correspondence, providing points contingent to selections of the green window after a correct response and selections of the red window after an error. Correspondence quickly increased and was maintained in subsequent baseline sessions. Correspondence training was effective to establish accurate reports of errors in these children.