Journal of Personality and Social Psychology 1997, Vol. 73, No. 6, 1380-1393 Copyright 1997 by the American Psychological Associalinn, Inc. 0022-3514/97/S3.OO Trait Self and True Self: Cross-Role Variation in the Big-Five Personality Traits and Its Relations With Psychological Authenticity and Subjective Well-Being Kennon M. Sheldon, Richard M. Ryan, Laird J. Rawsthorne, and Barbara Ilardi University of Rochester In 2 studies, college students evidenced differing levels of the "Big-Five" traits in different roles, supporting social-contextualist assumptions regarding trait expression. Supporting organismic theo- ries of personality, within-subject variations in the Big Five were predictable from variations in the degree of psychological authenticity felt in different roles. In addition, two concepts of self-integration or true selfhood were examined: 1 based on high consistency of trait profiles across roles (i.e., low- self-concept differentiation E. M. Donahue, R. W. Robins, B. W. Roberts, & O. P. John, 1993) and 1 based on high mean levels of authenticity felt across roles. The 2 self-integration measures were found to be independent predictors of psychological and physical well-being indicating that both self-consistency and psychological authenticity are vital for organized functioning and health. Admonitions to be true to oneself are as old as ancient philos- ophy and as perennial as moral lessons themselves. But what does it mean to be true to, or to act in accord with, oneself? At least two different answers can be gleaned from modern theories of personality, one that views people in terms of stable and enduring behavioral dispositions and one that views people in terms of a struggle for authenticity and self-expression. The first of these two perspectives has been offered by trait theorists, many of whom have recently galvanized around the "Big-Five" mode! of personality (McRae & John, 1992). In the trait view, people are assumed to have transcontextual (McCrae & Costa, 1984) personality dispositions that are highly stable over time, situations, and social roles. Not only do our traits characterize us, indeed, they may be "our very selves" (McCrae & Costa, 1994, p. 175) one implication of this view is that to be true to oneself is to behave in consistent accordance with one's own latent traits. Yet Big-Five theorists have been criticized precisely because of their focus on stability or consis- tency, to the seeming neglect of important social-contextual influences on personality (McAdams, 1992 Pervin, 1994 Ver- off, 1983). As Funder (1994) observed, it appears that trait theory "ignores individual differences in stability, ignores the way small changes in the environment can (sometimes) lead Lo large changes in personality, and fails to describe the mecha- nisms that promote stability and change" (p. 125). In short, the Big-Five model may not, as yet, provide a complete descrip- tion of personality. An alternative perspective, which stresses a more contextual Kennon M. Sheldon, Richard M. Ryan, Laird J. Rawsthorne, and Barbara Ilardi, Department of Psychology, University of Rochester. Correspondence concerning this article should be addressed to Ken- non M. Sheldon, who is now at Department of Psychology, University of Missouri, 203E McAlester Hall, Columbia, Missouri 65211. Elec- tronic mail may be sent via the Internet to psycks@showme. missouri.edu. and dynamic view of the person, is provided by organismic and existentially informed theories of personality (e.g., Deci & Ryan 1985 Rogers, 1963 Waterman, 1993). In these approaches, a central thesis is that people do not always act in accord with their self instead, they vary from situation to situation in the degree to which they contact and enact their true feelings and values. Roles and situations are assumed to differentially afford support for authentic self-expression and self-organized behav- iors, and some roles may foster false self-presentations, or de- partures from how one might ideally choose to be. According to this view, to be true to oneself within a role is to be able to behave in ways that feel personally expressive (Waterman, 1990), authentic (Ryan, 1993), or self-determined (Deci & Ryan, 1991). in this article, we investigate the relations between these seemingly disparate views of personality, by simultaneously ex- amining both the cross-role consistency of the Big-Five person- ality traits, and the construct of psychological authenticity, within five specific life roles. In so doing, we show a number of important differences, but also some important cross modula- tions and convergences between the two conceptions of personality. Cross-Role Variation in the Big-Five Traits The Big-Five model offers an integrative framework for per- sonality psychology (Costa & McCrae, 1995 Goldberg, 1993 McCrae & John, 1992). It focuses on a core set of behavioral traits���Extraversion, Neuroticism, Agreeableness, Conscien- tiousness, and Openness to Experience���and its proponents ar- gue that people can be understood by knowing how much they display each of these five traits in their lives. Workers in this tradition have amassed impressive evidence in support of the cross-cultural, cross-method, and temporal consistency of the Big-Five factors (McRae & John, 1992). As noted above, how- ever, the Big-Five model has been criticized for missing much of the action in personality, partly because it does not consider 1380

TRAIT SELF AND TRUE SELF 1381 that people might behave differently in different domains of life (Block, 1995 McAdams, 1992 Pervin, 1994). Because the aim of Big-Five assessment approaches is to locate people on underlying trait continua as reliably as possible, proponents of the model are prone to overlook cross-situational variation or to treat it as measurement error (Shadel & Cervone, 1993 Smith & Williams, 1992). The issue of situational differences in trait expression is thus peripheralized. Yet, it is apparent that people do behave differently in differ- ent roles and situations (Funder & Colvin, 1991) and that this variation is systematic (Roberts & Donahue, 1994) and mean- ingful (Ryan, 1995). Using one Big-Five trait as an example, it is reasonable to suppose that employee or student roles call for more conscientiousness than do other roles, such as those of friend or romantic partner. Such cross-situational variability seems natural and even expectable, given the diversity of the demands made by different roles and the diversity of the re- sources afforded within different roles. But this raises the ques- tion of the meaning and functional impact of role-to-role vari- ability in the expression of personality traits. Is such variability a sign of flexibility or of instability? And, do people vary in the extent to which they vary across roles? Cross-Role Variation as a Sign of Disorganization Recent research has shown that there are indeed individual differences in the degree of consistency people show across roles and, more, that this variation in variation covaries with health and adjustment outcomes. In one approach, Donahue, Robins, Roberts, and John (1993) created a self-concept differ- entiation (SCD) measure that indexes how distinctively people assess their different role-selves across representative sets of trait adjectives. Although these researchers were not specifically focused on the Big-Five traits, they found that people who mani- fest much inconsistency in trait profiles across different roles were lower on a variety of well-being indicators. Donahue et al. conceptualized high SCD as a state of nonintegration or self- fragmentation, in which the individual's functioning and well- being is impaired. Relatedly, in a series of within-subjects analy- ses, Roberts and Donahue (1994) showed that the relative sense of satisfaction a person feels within a particular role is positively associated with the degree of similarity evidenced between trait ratings made concerning that role and trait ratings made concern- ing "myself in general." The latter results suggest that when people vary away from their general or characteristic style within a given role, they tend to feel less content within that role. In short, there is good reason to suspect that highly discrepant modes of functioning create difficulties that negatively impact adjustment and well-being. William James aptly described such difficulties with a German term, Zerrissenheit, which literally means "tom-to-pieces hood" (cf. Lindbergh, 1955). A person who constantly remolds him or herself in line with role-related pressures or demands would seem to lack integrity and self- direction (Block, 1961) and might suffer accordingly. Note, however, that the logical converse of the self-fragmenta- tion idea proposed by Donahue et al. (1993) and Roberts and Donahue (1994) is an ideal of perfect self-consistency or invari- ance across roles. That is, a potential implication of the SCD research is that people are most integrated, and happiest, when they are exactly the same in every role. Taken as a prescription for how to attain well-being, Donahue et al.'s (1993) results suggest that people should strive to reduce their cross-role vari- ability and to behave in more uniform accordance with their latent general traits. However, the idea that perfect consistency represents perfect integration is theoretically problematic, be- cause one can also view such extreme invariance as a perfect lack of differentiation, potentially indicative of maladaptive ri- gidity and inflexibility (Funder & Colvin, 1991 Mischel, 1968). In this vein, Sheldon and Emmons (1995) have argued that differentiation and integration are distinct dimensions of person- ality organization and, thus, that the absence of one should not be mistaken for the presence of the other. Authenticity as a Sign of Organization Humanistic and organismic approaches to personality have taken a somewhat different approach in conceptualizing orga- nized functioning (e.g., Deci & Ryan, 1985 Rogers, 1963 Shel- don & Kasser, 1995). In these perspectives, integration is not defined as being consistent in one's life but, rather, as feeling authentic in one's life. Authenticity refers to behavior that is phenomenally experienced as being authored by the self (Wild, 1965) or internally caused (deCharms, 1968 Ryan, Deci, & Grolnick, 1995). People feet most authentic when they act with a full sense of choice and self-expression. We assume that such positive feelings give access to important internal resources, such as the ability to effectively regulate and maintain one's intentional states (Kuhl, 1986), the ability to process new infor- mation more deeply (Deci & Ryan, 1991), and the ability to think more creatively (Amabile, 1996). In short, felt authentic- ity is likely to be a good indicator of integration and organiza- tion, the fully functioning person (Rogers, 1963). According to the social-contextual assumptions of self-deter- mination theory (Deci & Ryan, 1985, 1991), not all situations or roles are conducive to choiceful and authentic behavior hence, there are expectable within-subject differences in the degree of authenticity felt within different behavioral domains (Ryan, 1995). Because felt authenticity is viewed as having dynamic impact on personality and behavior, people are ex- pected to manifest different behavioral styles in different roles, that is, to be inconsistent in their traits. Moreover, situations, roles, and domains that afford authenticity are viewed as facilita- tive of health and well-being, whereas situations, roles, and domains that stifle authenticity or autonomy are those that foster maladjustment and distress (Ryan, Sheldon, Kasser, & Deci, 1996 Sheldon, Ryan, & Reis, 1996). From this phenomenologi- cal perspective, it is not differentiation or variation per se that is indicative of fragmentation in personality but, rather, variation away from authenticity. The Present Research To coordinate these diverse viewpoints on the nature of per- sonality and of personality integration, we developed four work- ing hypotheses. First, we hypothesized that Big-Five trait scores, indeed, vary systematically across roles (Smith & Williams, 1992), in addi-