The effects of salient and non-salient stimuli on subjects' self report of a learning task

Abstract

This study addressed the hypotheses proposed by Nisbett and Wilson (1977) that 1) people do not have direct access in memory to the causes of their own behavior, but rather base their verbal reports of such causes on assumptions of plausible causal relationships, which are subject to various attributional biases, and 2) therefore, verbal reports on one’s own behaviour are not necessarily any more accurate than those of an observer provided with the same information regarding external stimuli (causes) and observable behaviours (effects). Sixty subjects participating in pairs in a learning experiment were assigned to one of 3 list conditions. The lists were designed to instill an expectation that either the color of the letters (black vs. coloured) or the order of the letters (meaningful vs. meaningless anagrams) is a potent variable in learning or that both may be potent. Subsequently, subjects participated in a second test condition as learners or observers, then made attributions of causal impact to both the letter arrangement ( a highly potent variable) of letter strings and/or colour (a highly salient but less potent variable). The results supported both an anti-introspectionist and pro-introspectionist position. On one hand, they supported the hypothesis that subjects would base their verbal reports on assumptions generalized (erroneously) from the first list. On the other hand, the results indicate that subjects attributions were not erroneously generalized from one phase to the other but subjects, in fact, made separate and accurate attributions to each learning phase. The hypothesis regarding the similarity of attributions of observer and learner subjects was not supported due largely to problems with the procedure . In particular, it is not clear whether observer subjects actually observed learners in the manner expected. Recommendations for future research are suggested.