Psychological Bulletin 1997, Vol. 121, No. 3, 371-394
Copyright 1997 by the American Psychological Association, Inc. 0033-2909/97/$3.00
Cynthia S. Symons
Blair T. Johnson
The Self-Reference Effect in Memory: A Meta-Analysis
In this review,the authors examine the basis for the mnemonic superiority that results from relating material to the self. A meta-analysisconfirms the expected self-referenceeffect (SRE) in memory, with self-referentencoding strategies yielding superior memory relative to both semantic and other- referent encoding strategies. Consistent with theory and research that suggest self-reference (SR) produces both organized and elaborate processing, the SRE was smaller (a) when SR is compared with other-reference(OR) ratherthan semanticencodingand (b) whenthe comparisontasks promote both organization and elaboration. Thus, the SRE appears to result primarily because the self is a well-developed and often-used construct that promotes elaboration and organization of encoded information.The authors discuss the implicationsof these and other findingsfor theories of the SRE and for future research.
Throughout the history of psychology, researchers have used the self as a central part of their explanations of various phenom- ena (see Banaji & Prentice, 1994; G. T. Greenwald & Pratkanis, 1984; and James, 1890). A large body of research suggests that the self-structure is unique, relative to other concepts (e.g., those about other people; see Kihlstrom et al., 1988; Markus, 1977; and Rogers, Kuiper, & Kirker, 1977), in its motivational and affective implications as well as in its structure and content. Social psychologists have long posited an important affective role for the self-concept (e.g., C. W. Sherif, Sherif, & Nebergall, 1965; M. Sherif & Cantril, 1947). More recently, appraisal theories of emotion have emphasized the phenomenological im- portance of the self in the interpretation of events and the re- suiting effect on emotions (Fiske & Taylor, 1991). From a moti- vational standpoint, examples of the self's pervasive influence abound. For example, the tendency to attribute another person's behavior to dispositional factors but one's own behavior to situational factors presumably occurs because the self dominates one's phenomenal perspective (Ross & Nisbett, 1991; Storms, 1973). Similarly, both self-serving biases and defense mecha- nisms have been attributed to self-protective or self-enhancing
Cynthia S. Symons, Department of Psychology,Houghton College; Blair T. Johnson, Department of Psychology,Syracuse University.
This article is based on Cynthia S. Symons's doctoral dissertation completedat SyracuseUniversityunderthe directionof BlairT.Johnson. The preparation of this article was facilitated by National Institute of Mental Health Grant MH01377-01.
We would like to thank Hung-YuLin, Beth Pavlik, and Kathy Paca- nowski for their assistance in gathering studies; KatherineBrundza and Diana Nichols for their help with some analyses; and Francis Bellezza, Kenneth Blick, John Mueller, and Glen Reeder for their kindness in sending unpublished materials. We would also like to express our ap- preciation to Harold Gelfand, KennethD. Levin, and Constantine Sedi- kides for valuable comments on earlier versions.
Correspondence concerning this article should be addressed to Cyn- thia S. Symons, Department of Psychology, Houghton College, Houghton, New York 14744. Electronic mail may be sent via Interact to email@example.com.
motives (cf. Fiske & Taylor, 1991; Maddi, 1989). The motiva- tional influence of the self in persuasion is evident when people resist persuasive appeals because of self-presentational concerns (Johnson & Eagly, 1989). Indeed, self-attention theory (Carver & Scheier, 1981) emphasizes that conformity of behav- ior to salient behavioral standards requires a focus on the self.
Given the breadth of interest in self-related phenomena and theories thereof (e.g., Cantor & Kihlstrom, 1987; Markus & Wurf, 1987), it is not surprising that researchers have more recently examined whether self-related processes invoke differ- ent memory stores (Klein & Loftus, 1993) and the extent to which the self-structure can be distinguished from structures about others (e.g., Aron, Aron, Tudor, & Nelson, 1991). In addition, research on self-schematicity (Markus, 1977; Mar- kus & Wurf, 1987) demonstrates that the content of self- schematic domains can have a wide variety of motivational, affective, and mnemonic consequences.
The focus of this article is on these purportedly unique mne- monic aspects of the self. Several researchers have argued that the self-structure in memory is unique relative to other concepts by virtue of its superior elaborative and organizational proper- ties as well as its frequent use in information processing (e.g., Kihlstrom et al., 1988; Maki & Carlson, 1993; Markus, 1977; Rogers et al., 1977; Singer & Kolligan, 1987). If the self indeed has superior elaborative and organizational properties, then in- formation actively related to the self should be better remem- bered than information that is processed in other ways (e.g., the relating of information to someone else or the processing of words for meaning). Researchers who initially obtained this pattern labeled the phenomenon the self-reference effect (SRE; Rogers et al., 1977). Although many subsequent studies found superior recall following self-reference (SR; e.g., Bellezza, 1984; Kuiper & Rogers, 1979; Maki & McCaul, 1985), other research suggest that the SRE was not so robust. Specifically, other kinds of non-SR processing appeared to promote memory as well as or better than SR (e.g., Bellezza & Hoyt, 1992; Keenan & Baillet, 1980; Klein & Kihlstrom, 1986; Lord, 1980). These conflicting study findings led Higgins and Bargh (1987) to conclude that "self-reference is neither necessary nor suffi-