|
Article Excerpt
Substantial evidence indicates that information is better remembered when it is processed in relation to the self than when it is related to someone else or when it is processed for general meaning (the so-called self-reference effect; SRE; e.g. Rogers, Kuiper, & Kirker, 1977). In their meta-analysis, Symons and Johnson (1997) found that the SRE occurs with highly significant regularity, although it is modulated by several factors, including the nature of the comparison task (other-reference versus semantic processing) and the degree of intimacy of the person used in the other-reference task. Specifically, a higher SRE was found in studies that compared self-reference (e.g. 'does this word describe you?') with a semantic task (e.g. 'does this word mean the same as XXX?') than in studies that compared self-reference with other-reference (e.g. 'does this word describe your mother?'). In addition, in the latter studies, the SRE was stronger when the person involved in the other-reference task was not someone with whom the subject was intimate (e.g. the subject's mother versus a celebrity; see Symons & Johnson, 1997).
Other studies suggest that the SRE also varies according to the affective valence of the stimuli. Indeed, it has been found that positive information (e.g. trait adjectives such as 'kind' or descriptions of behaviours such as 'I would take care of a sick friend for several days') is better recalled than negative information (e.g. trait adjectives such as 'dishonest' or descriptions of behaviours such as 'I would not pay back money that I owed to a friend') when it is processed in reference to the self, but not when it is processed in reference to another person or when it is processed for general meaning (Denny & Hunt, 1992; Kuiper & Derry, 1982; Sanz, 1996; Sedikides & Green, 2000, 2004). This influence of valence on memory for self-referent information is probably due to self-enhancement goals; that is, the motivation for most people to process positive rather than negative information about themselves (e.g. information indicating that they are competent, likable, attractive, etc.; Baumeister, 1998; Taylor & Brown, 1988). Specifically, when encountering new self-referent information, people may emphasize positive information and/or neglect negative information, thus enhancing the retention of positive as compared with negative information. In other words, the difference in memory for positive and negative self-referent information may result from differences in depth or elaboration of memory encoding. In agreement with this proposition, Sedikides and Green (2000; Experiment 3) found that increasing stimuli presentation time (from 2 to 8 seconds) enhanced recall of positive information, but not recall of negative information when this information was processed in reference to the self. The authors concluded that the memory effect is due, at least in part, to participants expending minimal processing resources while encoding negative information.
However, the memory difference between positive and negative self-referent information may also depend on processes occurring at retrieval. It has been proposed that, when retrieving information from memory, the self shapes retrieval cues in order to construct memories that are relevant to current goals of the individual (e.g. Conway & Pleydell-Pearce, 2000). By this means, access to negative self-referential information could be reduced, as this information may threaten the positive conception most people have of themselves (Baumeister, 1998; Taylor & Brown, 1988). If the difference in memory for positive and negative self-referent information is in part due to such a control on memory retrieval, then this difference should vary with retrieval conditions, including the amount and specificity of retrieval cues. Specifically, the difference should be more pronounced when there are few retrieval cues, as is the case in a free-recall task, than when retrieval cues are provided, as in a recognition task. Indeed, in a recall task, individuals must rely on strategic search processes (i.e. they have to elaborate their own retrieval cues) in order to access stored information (Koutstaal & Schacter, 1997; Norman & Schacter, 1996), thereby allowing the self to facilitate or, conversely, reduce the access to information (by shaping retrieval cues). In these conditions, positive self-referent information should be better remembered than negative self-referent information, which is precisely what has been found in studies that used a recall task to examine memory for positive and negative self-referent information (Denny & Hunt, 1992; Kuiper & Derry, 1982; Sanz, 1996; Sedikides & Green, 2000, 2004). In contrast, people have less need to elaborate their own retrieval cues in order to search for information stored in memory when they are confronted with a recognition task (Koutstaal & Schacter, 1997; Norman & Schacter, 1996). In this case, the self probably has less opportunity to exert control on the access to stored information because retrieval cues...
|
