THE TROUBLE WITH CHANGE; Self-Verification and Allegiance to the Self

Past approaches to the self have emphasized people'^ desire for positive evaluations I suggest that this emphasis overlooks another powerful and important motive, the desire for evaluations that venfy self-views Among people with negative self-views, this desirt for self-venfication can override the desire for positive evaluations For example, people with negative self-views seek relationship part tiers who view them negatively, elicit unfavorable evaluations from partners, and "see" more negativity in the reactions of others than actually there Although these self-venfication processes ordman impede progress in therapy, awareness of these processes can allo therapists to either circumvent them or actually use them in the se vice offostenng self-concept change For Ms W suffenng and victimizaUon were in some respects preferable to kindness and concem Ms W not only misperceived that Mr S was unfaithful but also resisted any informauon that contradicted her mispen,eption and ac uvely sought verification that he was unfaithful The better he treated her the more depressed and pessimistic she became (forj she was threatened by a canng and loving partner She accepted her past abuse as an appropnate reflection of her worth A challenge to this self-image was a challenge to how she adapted and coped with her vicumization (Widiger 1988 p 821) The responses of Ms W seem paradoxical because they defy the widespread conviction that all people possess a deep-seated need for praise and adulation It turns out that although people with negative self-views do at some level desire praise and adoration, they also want self-venficatwn in the form of evaluations that confirm and validate ir self-views This desire for self-confirmation appears to be an ;eedingly general one, one that shapes the lives of all of us, whether have high or low self-esteem In fact, it does not matter whether people's self-views are positive or negative, well-founded or misplaced, or based on something that happened dunng the previous year in the distant past Once people become confident of their selfviews, they rely on these self-views to predict the reactions of others, guide behavior, and to organize their conceptions of reality (e g , Mead, 1934) Because self-views must be stable to serve these vital functions, people work to venfy and confirm them (e g , Aronson, 1968, Secord & Backman, 1965, Swann, 1983, 1996) These selfvenfication stnvmgs may operate consciously or nonconsciously and may take several distinct forms FORMS OF SELF-VERIFICATION An especially important form of self-venfication occurs when people choose partners who see them as they see themselves, thereby :reating social environments that are likely to support their self-views n one study, for example, we asked people with positive and negative Address correspondence ilogy, Umversity of Texas @psy utexas edu VOL 8, NO 3, MAY 1997 o Wilham B Swann, Jr Department of Psyat AusUn Ausun, TX 78712, e-mail swann self-views whether they would prefer to interact with evaluator^ who had favorable or unfavorable unpressions of them As can be seen m Figure 1, people with positive self-views preferred favorable partners, and people with negative self-views preferred unfavorable partners (e g , Swann, Stein-Seroussi, & Giesler, 1992) More than a dozen replications in different laboratones usmg diverse methodologies have left little doubt that people with negative self-views seek unfavorable feedback and partners (e g , Hixon & Swann 1993, Robinson & Smith-Lovm 1992, Swann, Hixon, SteinSeroussi, & Gilbert, 1990, Swann, Pelham, & KruU, 1989, Swann, Wenzlaff, Krtill, & Pelham, 1992) Males and females display this propensity to an equal degree, irregardless of the degree to which the self-views are changeable or whether they are associated with specific qualities (mtelhgence, sociability, dommance) or global self-worth (self-esteem, depression) Similarly, people prefer to interact with self-venfymg partners even if presented with the alternative of participating in a different expenment (Swann, Wenzlaff, & Tafarodi, 1992) Finally, people are particulariy likely to seek seif-venfying evaluations if their self-views are extreme and finnly held (e g , Pelham & Swann, 1994, Swann, Pelham, & Chidester, 1988, Swann & Ely, 1984) Clinically depressed persons, for example, are more likely to seek negative evaluations than people with low self-esteem, presumably because depressives are thoroughly convinced that they are worthless (Giesler, Josephs, & Swann, 1996) People's efforts to venfy their negative self-views should not be confused with masochism For example, rather than savormg unfavorable evaluations (as one might expect masochists to do), people with negative self-views are intensely ambivalent about such evaluations In choosing a negative evaluator in one study (Swaim, StemSeroussi, & Giesler, 1992), one person with low esteem noted I like the [favorable] evaluauon but I am not sure that it is ah, correct maybe It sounds good but [the unfavorable evaluator] seems to know more about me So 1 II choose [the unfavorable evaluator] The thoughts that give nse to such ambivalence emerge sequentially Upon receiving and categorizing positive feedback, people are lmmeliately drawn to it, regardless of their self-views A preference for self-confirming feedback emerges later when people access their self•lewi, and compare these self-views to feedback (for a further discuslon of the mechanisms that seem to underlie self-venfication effects, see Swann, 1996, pp 55-69) The foregoing analysis implies that any procedure that prevents people from engaging in the companson process that gives nse to i-enfication stnvings should cause people with negative selfs to prefer favorable appraisals In support of this proposition, when my colleagues and I (Swann et al , 1990) had some people choose an mteracuon partner while they were depnved of cogmtive irces (by rushmg their decision or having them rehearse a phone number), we found that people with negative self-views were less inclined to self-venfy (i e , choose a partner who appraised them Copynght © 1997 Psychological Society PSYCHOLOGICAL SCIENCE Self-Venfication and Change Unfavorable Favorable Type of Evaluator Fig 1 Preferences for favorable versus unfavorable interaction partners among people with positive (pos ) versus negative (neg) selfconcepts The data on which this figure is based are from Swann, Stein-Seroussi, and Giesler (1992) unfavorably) Furthermore, after participants were no longer depnved of cognitive resources, they repudiated their earlier choices in favor of lf-venfymg ones Such findings suggest that when people with !gative self-views choose unfavorable feedback over favorable feedback. It IS because their desire for self-venfymg unfavorable feedback 'emdes their desire for favorable feedback ' Recent work indicates that if, despite their attempts to acquire self-venfymg feedback, people receive doses of self-discrepant feedback that cannot be readily dismissed, they become anxious (Pinel & Swann, 1996) In extreme cases, people may expenence what Kohut (1984) referred to as disintegration anxiety, a sinking feeling that something is tembly wrong accompanied by severe disonentation and ense of emptiness, incoherence, and worthlessness In light of the obvious aversiveness of disintegration anxiety, it is not surpnsing that people who receive disconfirming feedback take to counter it For example, when people suspect that others :ive them as being more or less likable than they perceive thems to be, they stnve to bnng the partners' evaluations into harmony with their self-views, even if (in the case of people with low self-esteem) this means lowenng the partners' evaluations (e g , in & Read, 1981, Study 2) Such compensatory activity, in tum, lzes people's self-views against self-discrepant feedback In one study, for instance, people who had an opportunity to resist a challenge to their self-views by "setting the evaluator straight' were less 1 The second stage that gives nse to self-venficaUon stnvings may later be ivemdden dunng a third stage m which people's responses are based on a cost-benefit analysis of charactenstics of the feedback, their self-views, and the social context Hence there appear to be at least three distinct phases in eople s reacuons to feedback an iniUal phase charactenzed by a preference 3r positive feedback a second phase charactenzed by a preference for conmem feedback, and a final phase dunng which people systemaUcally analyze le options available to them and behave so as to maximize their benefits and lmimize their costs For a further discussion, see Swann and Schroeder 1995) likely to expenence change in their self-views than those who had no opportunity to correct the evaluator (e g , Swann & Hill, 1982) Should the foregoing strategies fail to produce self-confirmmg social worlds, people may withdraw from the relationships in which they are receiving disconfirmmg feedback For example, if people wind up in mamages in which their spouses perceive them moi less) favorably than they perceive themselves, they become less mtimate with those spouses (Ritts & Stem, 1995, Swann, De La Ronde, & Hixon, 1994) If self-discrepant feedback is unavoidable, people may construct the illusion of self-confummg worlds by "seeing" more support for theuself-views than actually exists For example, just as people with positive self-views spend the longest time scrutmizing what someone says about them when they expect the remarks will be favorable, those with negative self-views spend the longest time scrutinizing whei they expect the remarks will be unfavorable (e g Swann & Read, 1981, Study 1) A parallel phenomenon emerges when researchers examine what people remember about the evaluations they rec Just as people with positive self-views remember more favorable than unfavorable statements that have been made about them, people with negative self-views remember more unfavorable than favorable statements (e g , Swann & Read, 1981, Study 3) And if these attentional and memonal processes are not enough tc insulate people against evaluations that challenge their self-views, people may nullify discrepant evaluations by selectively dismissing incongruent feedback For example, people express more confidence in the perceptiveness of evaluators whose appraisals confirm their self-conceptions (e g , Shrauger & Lund, 1975) 1 conjunction with the processes already outlined, such selective dismissal of challenging feedback may systemattcally skew people's perceptions of reality, encouraging them to conclude that their social worlds are far more supportive of theuself-views than is warranted Although these processes may stabilize people's self-views and foster feelings of coherence and predictability, they are also likely to impede lve psychological change IMPLICATIONS OF SELF-VERIFICATION PROCESSES FOR THERAPY Imagine a woman who seeks therapy in the hope of removing the self-doubt that has plagued her smce her youth Although the therapist may succeed in bnnging her to acknowledge and denve a feelmg of pnde from her strengths, she may also discover that these positive self-views are undone when she returns home to a husband who is itemptuous of her Such a scenano is not just hypothetical In one study, Predmore and I invited couples to the laboratory and seated srs in a room together Some intimates perceived their partner congruently and some perceived their partner less congniently At a ;y point in the procedure, we gave one member of each couple incongruent feedback When we later measured how much people's lews changed in the direction of the feedback, we found that participants were relatively impervious to the feedback if they were ining with an intimate who saw them congruently This tendency for congruent relationship partners to insulate one another against challenging feedback was equally apparent whether their self-views were sitive or negative (Swann & Predmore, 1985) Such evidence suggests an important addendum to Mark Twain's adage "A man cannot be comfortable without his own approval " To VOL 8 NO 3, MAY 1997 PSYCHOLOGICAL SCIENCE