Brown, J. D. (1991). Accuracy and bias in self knowledge. In C. R. Snyder & D. R. Forsyth (Eds.) Handbook of Social and Clinical Psychology: The Health Perspective (pp. 158-178). Elmsford, NY: Pergamon Press.
Jonathon D. Brown
Life is the art of being well-deceived; and in order that the deception may succeed it must be habitual and uninterrupted. (William Hazlitt, 1817)
Hazlitt's prescription for happiness notwithstanding, most observers of human nature have maintained that a-reliance on deception negatively affects well-being. Indeed, many theorists have regarded the ability to view the self and the world accurately as the sine qua non of psychological health. However intuitive and venerable this notion might be, a good deal of the recent research has failed to support the view that psychological well-being is characterized by the complete absence of distortion. Instead, accumulating evidence indicates that nearly all individuals, and particularly those who score highest on measures of psychological adjustment, exhibit a pervasive* tendency to view the self and their world in a more positive way than can realistically be justified (Lazarus, 1983; Sackeim, 1983; Taylor & Brown, 1988). Moreover, rather than undermining health, these illusions, as they have been referred to elsewhere (Taylor & Brown, 1988), appear to contribute to effective functioning, and also to foster well-being.
In this chapter, the self-enhancing nature of individuals' self-perceptions and the role that these perceptions play in promoting health will be considered. To begin, the structure and nature of the self-concept will be outlined. A number of theoretical perspectives that have emphasized the need for accurate self-knowledge will then be critically reviewed. Subsequently, the evidence that pertains to the illusory nature of individuals' self-conceptions will be presented, and the relation between self-enhancing illusions and adjustment will be discussed. Next, a number of mechanisms, both intrapersonal and social in origin, that enable individuals to maintain self-enhancing illusions will be considered. Finally, some additional issues, including the relation between illusions and defense mechanisms, and the practical implications of -the present analysis will be explored.
As it is used in this chapter, the term "the self refers to a cognitive structure that incorporates all of the ways in which a person characteristically answers the question, "Who am I." This usage corresponds to what James (1890) referred to as the self-as-known, or the "me." Throughout this paper, however, the self also is viewed as being actively involved in the processing of personal information. Thus, here the self is also taken to include self-as-knower, or the "I" in James's scheme. These dual aspects of the self are joined in what Markus (1977; Markus & Wurf, 1987) has labeled self-schemata.
An individual's answers to "Who am I” typically fall into three broad categories (Gordon, 1968; Rosenberg, 1979). First, there are one's physical attributes (e.g., I am a man"). Second, there are one's social identities. These identities are usually expressed as nouns and include the various formal and informal social roles that individuals occupy. They may refer to either ascribed (e.g., "I am a son') or attained (e.g., "I am a professor") roles or statuses. Finally, there are aspects of identity that are more individualistic in nature. These personal identities tend to be adjectival and include (a) one's perceived traits and dispositions (e.g., "I am impatient"); (b) one's presumed talents and abilities (e.g., "I am athletic"); and (c) one's attitudes, values, and interests (e.g., "I am an ardent environmentalist'). Together, physical characteristics and social and personal identities comprise the content of the self, or the self-concept.
When defined in this manner the self can be said to reflect individuals' subjective perceptions of who they are and of what they are like. These perceptions may or may not coincide with some objective measure. That is, a person's picture of the self (i.e., the self-concept) may or may not correspond closely to the "actual" self (i.e., the self as defined by some consensual or normative criteria). This is particularly true with respect to one's personal identities. That is, although there is generally little dispute regarding one's physical characteristics and social identities, one's personal identities are more open to interpretation. A student who earns high grades in her classes may nonetheless view herself as unintelligent and inadequate. Alternatively, an individual who conceives of himself as genteel and refined may actually be regarded by others as boorish and crude.
Given these possibilities, one may ask, What is the proper relation between the perceived self and the actual self? In particular, do well-adjusted, emotionally healthy individuals possess views of the self that closely parallel the actual self? Before turning to an empirical examination of this issue, it will be useful to first consider various theoretical perspectives on the matter. As will be seen, although they differ in many important respects, these perspectives share a common element: they all agree that individuals ought to and/or need to possess views of the self that are largely accurate and free of distortion.
Accuracy and Self-Knowledge
For centuries, philosophers, theologians, sages, and poets have enjoined us to look beneath the shroud of illusion and confront the self as it actually is. This injunction is embodied in the ancient dictum, "Know thyself," which has been attributed to Plato, Pythagoras, and Socrates, among others (see Gergen, 1971). The search for accurate self-knowledge appears to be encouraged for two reasons. First, self-understanding is championed on moral grounds. This perspective, which was initially advocated by early theologians and was later taken up by the existentialist philosophers (e.g., Nietzsche, Kierkegaard, Sartre), holds that individuals have an obligation to come to terms with their true nature; those who evade self-understanding are considered weak and cowardly and are perceived as living a depraved or purposeless existence (cf. Martin, 1985).
A second reason why individuals are urged to acquire veridical self-knowledge is for functional purposes. According to this view, successful commerce with the world requires a healthy measure of self-understanding. Indeed, legend has it that the need to "know thyself" was viewed with such supreme importance that this statement was inscribed on the Delphic oracle, a place where Kings allegedly came for guidance in matters of law, war, and the like.
In more recent years, the belief that accurate self-knowledge is essential for effective functioning has been embraced by psychiatrists, and personality and clinical psychologists. Somewhat paradoxically, with all of its emphasis on irrationality and distortion, Freudian theory represents one such view. Freud (1915/1957) believed that in their efforts to accommodate unacceptable instinctual impulses to the inexorable demands of the external world, individuals engage in self-deception in which their feelings, wishes, and desires are distorted or denied. The inevitable consequence of this self-deception is psychopathology. Consequently, the goal of psychoanalysis is to minimize distortion and reduce self-deception. Although Freud was never particularly sanguine about the possibility that individuals could achieve accurate self-knowledge, he was adamant in his belief that the price of self-deception was neurosis (see Sackeim, 1983, for a further discussion of these issues).
Other theorists also have endorsed the principle that psychological health depends on accurate self-knowledge. Jahoda (1958), for example, defined the mentally healthy person as one who is capable of perceiving the self as it actually is, without distorting one's perceptions to fit one's wishes. Similarly, Maslow (1950) wrote that healthy individuals are able to accept themselves and their own nature, with all of its discrepancies from their ideal image. Other psychologists (e.g., Allport, 1943; Erikson, 1950; Fromm, 1955; Haan, 1977; Menninger, 1963; Rogers, 1951) also have proposed that emotional well-being and accurate self-perception go hand in hand. Many of the so-called insight therapies these theorists helped develop are based on the principle that lasting therapeutic change can be achieved only when individuals come to view the self as it actually is.
In sum, though not the only view of mental health (e.g., Becker, 1973; Rank, 1936), many prominent theorists from the fields of psychiatry and personality and clinical psychology have asserted that the capacity to perceive the self, and the self's ability to act on the environment, without distortion are essential requirements for effective functioning. In some respects it is easy to see how such a perspective developed. Psychiatrists and clinical psychologists most often study abnormal behavior. Thus, the data base on which this view of mental health rests is comprised of individuals who experience a variety of pathologies. Some of these individuals suffer from severe psychoses, such as manic depression, schizophrenia, and the like. These conditions may be characterized by views of the self that are wholly without foundation in reality. For example, in the midst of a manic episode, an individual may believe that he has the power of a Napoleon or the charms of a Valentino. These beliefs provide compelling evidence that distorted views of the self can be unhealthy. It is tempting, therefore, to conclude that accurate perceptions of the self must be necessary for normal functioning. However, the logic underlying this inference process is faulty: The fact that mental illness is sometimes characterized by gross distortions of the self does not necessarily mean that mental health is characterized by the absence of distortion.
The broader issue here is that the study of abnormal behavior may not provide the best perspective on what normal behavior is like (cf. Baumeister, 1989; Snyder, 1989). Indeed, it seems arguable that if one wants to know what normal behavior is like, one ought to study normal individuals (i.e., individuals who are free from psychopathology). For the most part, this is the domain of social and health psychologists. Thus, one can ask, What does research in these disciplines reveal regarding the relation between the perceived self and the actual self'?
Although still incomplete in some respects, the portrait that emerges from decades of research in these and allied fields is not one that stresses the absolute accuracy of self-perceptions. Instead, normal individuals appear to consistently bias their self-perceptions in a self-enhancing direction Taylor & Brown, 1988). Three such tendencies are particularly evident: (a) individuals harbor unrealistically positive views of the self; (b) individuals exaggerate their ability to control environmental events, including events that are objectively uncontrollable; and (c) individuals are overly optimistic in that their beliefs about their future are brighter than can realistically be justified. In the sections that follow, the evidence pertaining to each of these illusions is reviewed.
Self-Enhancing Illusions
Unrealistically Positive Views of the Self
As previously noted, traditional conceptions of mental health maintain that well-adjusted individuals are aware of, and accepting of, both the positive and negative aspects of self. In contrast to this assertion, evidence indicates that most individuals conceive of the self in terms that are overwhelmingly more positive than negative. That is, nearly all individuals believe that they possess far more positively valued than negatively valued characteristics (Brown, 1986; Brown & Gallagher, 1990; Dunning, Meyerowitz, & Holzberg, 1989).
By itself, this imbalance does not provide evidence that such views are unrealistic or illusory. After all, most individuals may possess more positive than negative qualities. Evidence for the illusory nature of such beliefs does exist, however, First, nearly all individuals show a pervasive tendency to see the self as "better" than others. They judge positive attributes to be more descriptive of the self than of the average person, but see negative attributes as less descriptive of the self than of the average person (Brown, 1986; Brown & Gallagher, 1990; Dunning et al., 1989). Moreover, far from being limited in scope, this self-other bias (Brown, 1986) has been documented for a wide range of traits and abilities (Alicke, 1985; Dunning et al., 1989). Insofar as it is logically impossible for most people to be better than the average person, these highly skewed, positive views of the self provide evidence for their unrealistic and illusory nature.
The overly positive manner in which individuals view the self also extends to appraisals of their friends and intimates. Specifically, although the tendency to see the self as better than others is attenuated when the others being evaluated are close friends or relatives, there exists a corresponding tendency for individuals to see their companions as better than average (Brown, 1986). Moreover, these effects at the individual level also occur at the group level. Even under the most minimal of social conditions, there is a pervasive tendency for individuals to see their own group as better than other groups (Brown, Collins, & Schmidt, 1988; for a review, see Tajfel & Turner, 1986). Thus, although research also demonstrates a general positivity bias in person-perception (Sears, 1983), individuals are inclined to appraise their friends, family, and own group members in far more positive and less negative terms than they appraise most other people. Again, because it is logically impossible for most people's associates to be better than average, this tendency may be regarded as illusory.
A second source of evidence pertaining to the illusory nature of positive self-perceptions comes from investigations that have compared self-ratings with judgments made by observers. Lewinsohn, Mischel, Chaplin, and Barton (1980) had observers watch subjects complete a group interaction task. Observers then rated each subject along a number of personality dimensions (e.g., friendly, warm, assertive); subjects also rated themselves along each of the attributes. The results showed that self-ratings were significantly more positive than were the observers' ratings. In other words, individuals saw themselves in more flattering terms than they were seen by others (see also Gotlib & Meltzer, 1987).
In sum, the perception of self that most individuals hold is not as well balanced as traditional models of mental health suggest. Rather than being attentive to both the favorable and unfavorable aspects of self, normal individuals appear to be very cognizant of their strengths and assets and considerably less aware of their weaknesses and faults. Evidence that these flattering self-portrayals are illusory comes from studies that have found that (a) most individuals see themselves (and their intimates) as better than the average person, and (b) most individuals see themselves as better than they are seen by others (Taylor & Brown, 1988).
Illusions of Control
Illusory beliefs regarding the positivity of the self are accompanied by a second illusion; namely, an exaggerated belief in one's ability to control the environment. Several lines of research from the areas of human learning, social, and clinical psychology suggest that individuals often believe that they can control events, even when the events are entirely determined by chance (for a review, see Abramson & Alloy, 1980). Moreover, this illusion of control (Langer, 1975) is particularly evident when desirable outcomes are obtained (Alloy & Abramson, 1979; Greenwald, 1980). Thus, although accurate knowledge of the relation between one's actions and environmental outcomes would seem to be essential for effective functioning, a good deal of evidence suggests that people overestimate the degree to which their actions produce environmental events, especially when these events are positive in valence.
Unrealistic Optimism
The belief that one is good and that one can control events gives rise to a third self-enhancing illusion. This illusion may be termed unrealistic optimism. Simply put, most people believe that their future will be brighter than base rate data can justify. People estimate the likelihood that they will experience a wide variety of pleasant events, such as liking their first job, getting a good salary, or having a gifted child, as higher than those of their peers (Weinstein, 1980). Conversely, when asked their chances of experiencing a wide variety of negative events, including having an automobile accident (Robertson, 1977), being a crime victim (Perloff & Fetzer, 1986), having trouble finding a job (Weinstein, 1980), or becoming ill (Perloff & Fetzer, 1986; Weinstein, 1982) or depressed (Kuiper & MacDonald, 1982), most people believe that they are less likely than their peers to experience such negative events. Insofar as everyone's future cannot logically be rosier than average, the extreme optimism individuals display appears to be illusory.
It should be noted that beliefs in personal control and perceptions of the future need not always involve self-referent cognitions. However, many times they do. For example, with respect to perceptions of control, many theorists have argued that a sense of agency is integral to the self-concept (White, 1959). When subjects overestimate their control over objectively uncontrollable outcomes, they seem tacitly to be saying, "I am efficacious." Similarly, although beliefs about the future do not always center around the self, Markus and Nurius' (1986) work on possible selves suggests that individuals' notions of the future are often wedded to their thoughts about the self. For these reasons, then, like overly positive views of the self, illusions of control and unrealistic optimism are treated as self-referent cognitions in the present paper.
Self-Enhancing Illusions and Psychological Adjustment
The previous discussion documents that most individuals harbor unrealistically positive views of the self, their ability to control events in their lives, and their future. These findings seem to be at odds with the claim that accurate self-perceptions are necessary for effective functioning. However, evidence relating illusions to adjustment has yet to be presented. If, for instance, the illusions documented above were found to be more prevalent among individuals who are psychologically distressed than among those who score high on measures of psychological adjustment, traditional notions of mental health would be upheld.
Illusions and Depression
A useful point of departure for addressing this issue is to examine the relation between illusions and the depressive disorders. Although not the only form of psychopathology, depression is far and away the most common. According to some estimates, depression accounts for 75 % of all psychiatric hospitalizations and at any given time, 15% of the adult population suffers from depressive symptoms (cited in Abramson & Martin, 1981). If self-enhancing illusions are detrimental to psychological health, they should be positively linked to depression.
In fact, an examination of the association between illusions and depression suggests that just the opposite is true. A good deal of research subsumed under the rubric of "depressive realism" has found that self-enhancing illusions are relatively absent among depressed people (Alloy & Abramson, 1988). Specifically, compared with nondepressed people, those who are moderately or severely depressed (a) are relatively more attentive to the self's positive and negative qualities (Kuiper & MacDonald, 1982); (b) display greater congruence between self-evaluations and evaluations of others (Ahrens, Zeiss, & Kanfer, 1988; Kuiper & MacDonald, 1982); and (c) offer self-appraisals that coincide more with those of objective observers (e.g., Lewinsohn et al., 1980). Additionally, they arc less likely to fall prey to an illusion of control (Alloy & Abramson, 1979; Golin, Terrell, & Johnson, 1977; Golin, Terrell, Weitz, & Drost, 1979) and are less apt to display unrealistic optimism about the future (Alloy & Ahrens, 1987; Pyszczynski, Holt, & Greenberg, 1987). In short, relative to nondepressed people, depressives tend to show more balanced and/or accurate perceptions of the self, the degree to which their responses control environmental outcomes, and their future. These findings, which are somewhat inconsistent with traditional cognitive models of depression (e.g., Beck's (19671 notion of the negative cognitive triad), pose a challenge to the claim that accurate self-knowledge is linked to psychological health.
Other areas of research offer additional evidence that accurate self-views may be linked to depression and other forms of psychological distress (i.e., neuroticism, anxiety, low self-esteem). The Self-Deception Questionnaire (Sackeim & Gur, 1979) measures the degree to which individuals typically deny psychologically threatening but universally true statements (e.g., "Do you ever feel guilty?"). High scores on this scale have been found to be negatively related to depression and neuroticism (Roth & *Ingram, 1985; Sackeim, 1983). The fact that individuals who are most prone to engage in self-deception also score lowest on some measures of psychopathology further suggests that accurate self-knowledge may not be essential for mental health.
Conceptually similar findings come from research on the correlates of private self-consciousness (Fenigstein, Scheier, & Buss, 1975). Private self-consciousness refers to the degree to which a person characteristically attends to the private, covert aspects of the self (e.g., "I'm always trying to figure myself out"). People who are high in private self-consciousness (i.e., those who are generally attentive to these aspects of the self) have been shown to possess more detailed and accurate self-knowledge than those who are low in private self-consciousness (Franzoi, 1983; Turner 1978). Research also has found that private self-consciousness is positively related to depression (Ingram & Smith, 1984; Smith & Greenberg, 1981; Smith, Ingram, & Roth, 1985). In other words, those who "know" themselves best score highest on measures of affective distress.
In sum, self-enhancing illusions are more characteristic of psychologically healthy individuals (i.e., those who are nondepressed, low in neuroticism) than they are of individuals who are less well adjusted. In some cases, the absence of self-enhancing illusions on the part of distressed individuals may reflect undue pessimism and self-deprecation. Nonetheless, there is scant evidence that psychologically healthy individuals are less biased or more accurate in their self-perceptions than are those who suffer from some of the most common forms of psychological distress. Consequently, there is little evidence that evenhanded and realistic perceptions of self are essential for mental health.
Illusions and Achievement
Although emotional well-being is certainly an important component of mental health, it is not the only element of psychological adjustment. Another important constituent of adjustment is the capacity for creative and productive work. Interestingly, illusions also have been linked to success in the achievement domain. A number of studies have shown that individuals who approach intellectual or creative tasks with a belief in their ability, an attitude that such tasks can be mastered, -and a high expectancy of success perform better than do those who display an absence of these tendencies (e.g., Bandura & Schunk, 1981; Baumeister, Hamilton, & Tice, 1985; Feather, 1966). Moreover, these effects occur even when objective assessments of ability indicate that those
who exhibit positive self-relevant cognitions are no more able than are those who display an absence of such tendencies (Dweck & Leggett, 1988). Thus, under some conditions, the belief that one can succeed is a better predictor of success than is actual ability level.
The positive relation between illusions and achievement appears to be most evident for tasks of moderate or high difficulty and is mediated by several factors. First, individuals with high self-perceptions of ability adopt more efficient problem-solving strategies than do those who harbor doubts about their ability to succeed (Bandura, 1989; Dweck & Leggett, 1988). Second, whereas individuals who are confident in their abilities are able to keep their attention focused on the task at hand, the attention of those with low self-perceptions of ability tends to wander toward task-irrelevant cognitions that serve to undermine performance (Sarason, 1975). Third, individuals with positive self-views and a belief in their abilities tend to work harder and persist longer at difficult tasks than do those with negative self-views; consequently, they perform better, particularly after prior failure (Bandura, 1989; Felson, 1984; Shrauger & Sorman, 1977). Because nearly all important goals in life entail periodic setbacks and thus require that obstacles to goal attainment be surmounted, it follows that individuals with positive self-views will reach higher levels of achievement in the real world. Although the evidence here is sparse, an investigation by Mortimer and Lorence (1979) supports this conjecture (see also Bachman & O'Malley, 1977). Finally, it is important to note that perseverance can sometimes be maladaptive, as when one endlessly persists at a task that is truly intractable (Janoff-Bulman & Brickman, 1982). Here again, however, the advantage may lie with those who possess positive views of the self. Individuals with positive self-conceptions appear to be most sensitive to the conditions under which persistence will and will not pay dividends (Janoff-Bulman & Brickman, 1982; McFarlin, 1985; Sandelands, Brockner, & Glynn, 1988; (see also Baumeister & Tice, 1985; McFarlin, Baumeister, & Blascovich, 1984]).
The associations between illusions and achievement are particularly interesting because they contradict the widespread belief that individuals need to know their true ability level in order to maximize their outcomes in life. Although drastically overestimating one's ability can certainly beget disappointment and failure, performance does not solely (or even predominantly) depend on ability. Instead, achievement -outcomes typically depend on a host of other factors, including the particular strategies one uses, and how hard and how long one tries. As just noted, beliefs in one's ability have been shown to be positively related to all of these factors. That is, in achievement-related situations, high self-perceptions of ability are linked to the use of more effective strategies and to greater motivation and persistence (Bandura, 1989). Although the presence of these factors does not guarantee that success will be realized, their existence almost assuredly results in a higher level of goal attainment than would be possible in their absence. Thus, positive perceptions of one's ability, even if somewhat illusory, may actually enable individuals to maximize their likelihood of attaining important goals.
Summary
In this section the fink between self-enhancing illusions and psychological adjustment has been reviewed. In some cases further research is needed. But the bulk of the evidence indicates that self-enhancing illusions are associated with superior psychological functioning. Individuals who possess positive views of the self, a steadfast belief in their ability to control events, and a positive view of the future are less apt to be depressed than are those who fail to exhibit such beliefs. They also are more apt to succeed at achievement-related activities. Thus, far from undermining psychological health, the evidence suggests that self-enhancing illusions are positively related to well-being and adjustment.
MAINTAINING ILLUSIONS
As we have documented the prevalence of self-enhancing illusions and considered their relation to psychological health, many readers may well be wondering at this point how individuals are able to harbor illusions with apparent impunity. Even in the most benign of worlds, it would seem that people would occasionally have to put their self-enhancing beliefs to a test. In other words, doesn't reality intrude on one's private musings, at least from time to time?
At the risk of being glib, the first answer to this question is "it depends." More specifically, it depends on how we define reality. In his analysis of this issue, Watzlawick (1976) distinguished between two aspects of the term. The first may be
called perceptive reality; this facet refers to the sensory perception of the objective and tangible property of objects and events. The second may be called interpretive reality; this facet refers to the subjective meaning individuals ascribe to these objects and events. The critical difference between these terms is their relative verifiability. Although our perceptions of physical reality generally are subject to confirmation, our interpretations of reality remain very much in the realm of subjectivity. To illustrate, whether I am writing a paragraph can be verified with relative ease; whether I am writing an exceptionally insightful paragraph, however, is very much open to interpretation. Thus, when speaking of the inevitable intrusion of reality, one must distinguish between perceptive reality and interpretive reality. Whereas individuals are constrained in their perceptions of physical reality, they are relatively free to construct their own interpretive reality.
For the most part, the illusions we have considered involve matters of interpretive reality, not ones of perceptive reality. Notions of how compassionate, kind, and courageous we are, for example, are not subject to verification. Objective measures of compassion do not exist; standardized tests of kindness await development. Nor can one simply refer to a class of behaviors to define a person's standing on attributes like these. For some, race car driving may represent the essence of courage; for others, it constitutes the epitome of foolishness and immaturity. Neither of these interpretations can be considered inaccurate because there is no absolute standard of accuracy against which they can be measured. Given the inherently subjective nature of traits such as these, individuals have more or less free rein to define their standing on many, if not most, personal attributes (Dunning et al., 1989). In a large sense, then, the indeterminate nature of social reality promotes the use of self-enhancing illusions.
One implication of the preceding analysis is that self-enhancing illusions ought to be more evident for relatively ambiguous attributes than for attributes that are more easily subject to verification. An investigation by Felson (1981) supports this conjecture. Felson had college football players rate themselves on a number of attributes relevant to performance in football; the players' coaches also evaluated the players along these dimensions. Some of the abilities (e.g., speed, size) were considered relatively unambiguous and verifiable insofar as clear standards exist for assessing one's standing on these attributes (speed can be measured, for instance). The remaining abilities (e.g., mental toughness, football sense) were deemed to be more ambiguous and subjective in nature insofar as one's standing on these measures cannot easily be determined. The prediction was that the player's estimates would be more likely to exceed the coaches' on the ambiguous attributes than on the unambiguous attributes. This prediction was confirmed (see also Dunning et al., 1989). Thus, people's self-appraisals may remain relatively close to the "data" with respect to attributes that are easily subject to verification, while their most aggrandizing self-assessments are reserved for attributes that are more subjective in nature.
But what about those attributes that are capable of objective verification? How, for example, are the majority of people able to cling to the belief that they are above average in intelligence when standardized test scores and the like seem readily available to potentially convince them otherwise? Research suggests that individuals are able to sustain favorable self-perceptions like these by closely adhering to three general strategies. These involve (a) behavioral strategies that minimize the likelihood that negative self-relevant feedback will be encountered; (b) cognitive strategies that lessen the probability that negative feedback, when encountered, will be registered or perceived as implicating the self; and (c) "damage control" strategies that reduce the chances that negative feedback, when encountered and acknowledged, will impact negatively on overall feelings of self-worth. The nature of these strategies is explored in the ensuing sections.
Behavioral Strategies for Avoiding
Negative Feedback
The first class of strategies to be discussed comprises behavioral attempts to ensure exposure to negative feedback will be minimal. This goal can be achieved either (a) by selective exposure to positive feedback or (b) by taking active efforts to ensure that the causes of negative outcomes are ambiguous.
Selective Exposure to Favorable Feedback
One way individuals would be able to sustain their self-enhancing beliefs would be to arrange matters so that negative self-relevant feedback is never encountered. However, completely insulating oneself from receiving negative feedback is doomed to be maladaptive. An individual who remained entirely oblivious to his lack of ability in some domain would be condemned to experience repeated failure in that aspect of life. A more modest, but infinitely more adaptive, strategy would be to approach positive self-relevant information more vigorously than negative self-relevant information. In this manner, the preponderance of the feedback one received would be positive, but negative feedback, though not actively sought, would still be encountered from time to time.
Evidence supporting just such a biased pattern of information-seeking behavior has been uncovered (e.g., Brown, in press; Frey & Stahlberg, 1986; Pyszczynski, Greenberg, & LaPrelle, 1985). For example, Brown (in press; Study 1) first gave subjects success, failure, or no feedback on a pretest. Subsequently, they were given the opportunity to learn more about their ability level by choosing to work on one of two tests during the second part of the experiment. One of the tests allegedly provided subjects with immediate performance feedback, whereas feedback was allegedly unavailable for a second task. The prediction was that individuals would seek self-relevant feedback less actively when they expected to receive unfavorable as opposed to favorable information. As anticipated, the tendency to choose the test yielding performance feedback was significantly less pronounced among subjects who anticipated finding out that their ability was low (i.e., those who had experienced prior failure) than among those who expected to succeed. In a subsequent study, this ambivalence toward gaining negative information concerning one's abilities was not found to depend on differential perceptions of test validity or negative mood states. Nor did this information-seeking strategy merely represent a desire to maintain a public image of competency in the eyes of others: Further investigation found that this tendency was more prevalent under private conditions than under public conditions. Collectively, the findings suggest that individuals may actively approach feedback about their abilities when they expect it to be positive but largely refrain from seeking feedback about their abilities when they expect it to be negative. This pattern of information-seeking ensures that most of the information they receive about their abilities will be favorable.
At first blush, these findings seem inconsistent with prior research regarding self-evaluation of abilities in achievement settings. Several investigations have found that when subjects are offered a choice between tasks that vary in their informational value, most subjects choose tasks that provide them with the most ability-relevant information (for a review, see Trope 1986). These findings have been interpreted to indicate that individuals seek information about their abilities without regard to whether this information is apt to be positive or negative (Trope, 1986). However, most prior studies have not given sufficient attention to the fact that the vast majority of subjects believe that they have high ability, and thus expect to succeed at the target task. Consequently, like those subjects who recently experienced success in the Brown (in press) studies, the preference they have shown for tasks of high informational value is easily understood as representing a desire to confirm an image of competency and gain additional favorable information about the self.
Two additional points about selective exposure to positive feedback are worth noting. First, avoidance of negative feedback is not always obvious or deliberate (Greenwald, 1988). Oftentimes individuals may believe that they possess some ability or talent but are not sure. They can evade finding out whether the ability is truly present or not by avoiding situations that call for its display (Shrauger, 1982). For instance, suppose that in the privacy of his mind (or shower) an individual believes that his singing voice is second only to Sinatra's. By judiciously avoiding situations that call for public singing, he would never have to put this belief to a test. As a result, he would forever be free to cling to the belief that he is a spellbinding vocalist.
Second, it is important to acknowledge that individuals do often seek out purely diagnostic feedback on aspects of the self that are modifiable'. Many of us routinely send our manuscripts to others for comments, or-ask our colleagues for help in polishing a talk before delivering an important colloquium. Although, to some extent, this practice probably stems from an expectation that the remarks we receive will be favorable, it also reflects a genuine desire to improve the quality of our work. But seeking feedback about the products of one's ability is not the same as seeking feedback about one's ability per se. Whereas one's work is subject to modification and improvement, one's abilities are relatively fixed. Consequently, though we may ask for feedback on what we have produced, few of us ever ask our colleagues to tell us whether they truly think we possess the intellectual ability to make a contribution to the field, whether we have any innate ability as a writer, and so forth.
Self-Handicapping Strategies
Of course, individuals do not always avoid diagnostic situations; nor can they. A student who is doing poorly in a course, for example, typically does not have the luxury of choosing whether or not to take the final exam if the class is one that is required for his major. Although in such situations a negative outcome may be inevitable, individuals may be able to control the degree to which the negative outcome implicates central aspects of the self. They may do so, according to Berglas and Jones (1978), by engaging in self-handicapping strategies. Interestingly, these strategies entail erecting self-induced barriers to performance in order to ambiguate the cause of an anticipated negative outcome. The student who fears that his test performance will be less than exemplary may sabotage his performance by not studying for the impending exam. If he receives the poor grade he expects, lack of effort provides a ready cause of failure; the extent to which the negative outcome reveals low ability is therefore obfuscated. Mom-over, if by chance he should succeed, lack of effort provides persuasive evidence of high ability. After all, who but a veritable genius could succeed when saddled with the impediment of insufficient preparation?
To test whether individuals engage in self-handicapping behaviors when failure is anticipated, Berglas and Jones (1978) first led some male subjects to believe that they were likely to succeed on an upcoming test; others were given sufficient mason to believe that future success was unlikely. All subjects were then told that the second part of the experiment involved testing the effects of two new drugs on test performance. One of the drugs purportedly facilitated test performance, the other supposedly impaired performance. The subjects were then given a choice as to which drug they wished to ingest. Consistent with the notion that individuals will take active efforts to sabotage their performance when future success is improbable, subjects who believed that success was unlikely were most apt to select the performance-inhibiting drug. These findings suggest that in at least some situations where negative feedback is unavoidable, individuals actively ensure that failure does not implicate valued aspects of the self (i.e., low ability). In this manner, they are able to cling to an image of competency even when negative outcomes are encountered.
Extensions to Social Behavior
Thus far, only intrapersonal information -seek ing strategies have been discussed. However, each of the tendencies reviewed above has parallels in the interpersonal domain. With respect to selective exposure to favorable feedback, research indicates that individuals gravitate toward social relationships in which their interaction partners view them in a manner that is consistent with how they see themselves (Swann, 1987). Because most people have positive self-views, most choose to associate with others who also view them positively. This ensures that most of the interpersonal feedback that they receive will be favorable (cf. Rosenberg, 1919). Even in situations where people must interact with others who hold views of them that are discrepant from their own, they may be able to orchestrate the interaction so that their partners are virtually forced to confirm their self-views (Swann & Hill, 1982). Again, because most people hold favorable views of the self, these constraining interaction strategies mean that predominantly positive feedback about the self will be encountered. Finally, self-handicapping strategies are relevant to the affiliative domain. A suitor who fears rejection may wait until the last minute to call his intended for a date. If his advances meet with the rejection he anticipates, he can attribute the negative outcome to his failure to call early enough, rather than to more central aspects of the self (e.g., an obnoxious personality). Other social factors that are not specifically related to information-seeking patterns also maximize the likelihood that individuals receive mostly positive feedback from their interaction partners. For instance, although people are generally reluctant to give interpersonal feedback, they are especially disinclined to do so when feedback is negative (e.g., Blumberg, 1972; Tesser & Rosen, 1975). Normative standards like these ensure that most of the feedback we receive from others is favorable.
Cognitive Strategies for Coping with Negative Feedback
Direct avoidance of negative feedback and self-handicapping strategies may be considered proactive behavioral tactics that individuals deploy to lessen the likelihood that negative feedback about the self will be encountered. However, because these strategies are used only when negative feedback is anticipated, they do not guarantee that self-deflating feedback can be evaded entirely. Oftentimes negative feedback is unexpected, as when one anticipates victory but tastes defeat. Under such conditions, do individuals passively resign themselves to receiving negative feedback?
Clearly not! Instead, they rely on a series of cognitive strategies to either color the feedback they receive or, at the very least, cushion the adverse impact of negative feedback. These strategies involve selectivity in how individuals interpret, attend to, remember, and explain evaluative feedback. The evidence relating to each of these tendencies is briefly summarized below.
Selective Interpretation
In many evaluative situations the feedback one receives is somewhat ambiguous. Suppose, after a good night kiss on the doorstep, one's date remarks, "Well, I can say in all honesty that I've never been kissed like that before." Whether this is the grandest of compliments or a monumental put-down is not entirely clear. Given this latitude, individuals are free to interpret ambiguous feedback in a benign manner if they are so inclined. An investigation by Dykman and Volpicelli (1983) suggests that many individuals are so inclined. On each of 40 trials, these researchers gave subjects either success, failure, or ambiguous feedback regarding their task performance. Through the course of the experiment, nondepressed subjects increasingly came to interpret ambiguous feedback as positive feedback. Interestingly, this bias was not present among depressed subjects.
Selective Attention
Even in situations where the feedback one receives is relatively unambiguous, there generally exists a mixture of good news and bad news. For example, even manuscripts that meet with rejection typically come back with reviews that include perfunctory plaudits regarding the paper's organization or writing style and obligatory remarks concerning the potential for publication in another journal. By selectively attending to these favorable comments, individuals can cognitively "snatch victory from the jaws of defeat."
Somewhat surprisingly, there is little evidence directly testing whether individuals selectively attend to positive feedback (but see Mischel, Ebbesen, & Zeiss, 1973; Roth & Rehm, 1980, for exceptions). There is, however, a good deal of evidence that individuals selectively attend to information that is congruent with their expectations (for a review, see Fiske & Taylor, 1984), and this includes expectancies that implicate the self (Swann & Read, 1981). Considering that most people have positive self-views and expect positive outcomes, these general findings suggest that most people selectively attend to positive self-relevant information.
Selective Memory
In addition to biases in interpretation and attention, individuals also can influence the nature of the feedback they receive by differentially recalling performance feedback. If after giving a presentation one remembers all of the compliments one received but few of the criticisms, one might easily conclude that the presentation was a resounding success (Gilbert & Cooper, 1985). In line with this notion, research indicates that most people, and particularly those who are nondepressed or high in self-esteem, exhibit better memory for positive feedback than for negative feedback (Johnson, Petzel, Hartney, & Morgan, 1983; Mischel, Ebbesen, & Zeiss, 1976; for a review, see Greenwald, 1980).
Selective Attributions
A final strategy for limiting negative feedback is to acknowledge its existence but deny its implications. One of the most reliable findings in social psychology over the last 20 years has been a pervasive tendency on the part of most individuals to accept greater personal responsibility for success than for failure (for reviews, see Snyder, Higgins, & Stucky, 1983; Taylor & Brown, 1988; Zuckerman, 1979). In particular, whereas positive outcomes are typically attributed to stable, central aspects of the self (e.g., "I received a high test grade because I am smart"), negative outcomes tend to be attributed to external factors (e.g., "I received a low test grade because the test was unclear') or, at the very least, less stable and/or central aspects of the self (e.g., "I received a low test grade because I studied the wrong material"). By admitting that only positive outcomes are due to one's traits, abilities, and/or dispositions, individuals are able to hold onto their self-enhancing beliefs even when confronted with failure.
How are such self-enhancing explanations for events derived? As noted elsewhere (Kunda, 1987; Pyszczynski & Greenberg, 1987), one possibility is that individuals bypass all semblance of logical information processing and decide on a cause that "feels good." However, such a process would be problematic for at least two reasons. First, most individuals like to think of themselves as being rational, thoughtful, and usually quite logical. To completely disregard all rules of logic when making causal judgments would threaten this image. Moreover, explanations for events that are wholly without logical underpinnings would seemingly be easily discredited and vulnerable to disconfirmation. For these reasons, then, it is unlikely that self-enhancing causal explanations for events are reached in a careless, haphazard way.
Instead, self-serving attributions are apt to be generated by a process that, at least to the attributor, seems logical. One attempt to specify how this process might proceed has been offered by Pyszczynski and Greenberg (1987). Their model begins by assuming a general sequence of the attribution process of the following form. After an event occurs, a plausible causal hypothesis is generated. Inference rules needed for testing the hypothesis are then settled on. Subsequently, data relevant to testing the hypothesis are gathered, and the validity of the data evaluated. Finally, the data are weighted and integrated and a final causal judgment is reached.
Pyszczynski and Greenberg (1987) assume that a desire to reach a self-enhancing conclusion can potentially exert an important influence at each of these steps. To illustrate, reconsider the student who has done poorly on an important exam. Initially, the student may generate a self-serving causal hypothesis (cf. Kunda, 1987). She may decide that the poor performance was probably due to ambiguous test questions rather than to her own lack of ability. At this point, she might settle on an inference rule that is especially congenial to her self-serving hypothesis. Perhaps she concludes that in order to properly test her hypothesis, she need only determine whether some of her fellow students also found the questions to be lacking in clarity. When gathering data relevant to testing this proposition, she might then be prone to sample from the population in such a way that her hypothesis is apt to be supported. For example, she might query only students who did at least as poorly, if not worse, on the exam as she did (Pyszczynski et al., 1985). If these students also found the test questions to be vague and equivocal, her hypothesis would seemingly have received support. In the event that any evidence inconsistent with the hypothesis is encountered, it can be dismissed as invalid or, at the very least, less relevant. For instance, if another student who did poorly did not find the questions confusing, the attributor may dismiss that student's perceptions as atypical and aberrant (e.g., "He's so out of it, he probably didn't even read the questions!'). By adhering to such a strategy, the attributor is able to cling to the belief that her conclusion regarding the causal role of the ambiguous test questions is fully justified on the basis of the available evidence. The fact that this process violates the dictates of formal logic and deviates from normatively correct rules of hypothesis-testing is apt to be of little concern; phenomenologically, the process is likely to be perceived as logical, thorough, and appropriate. Thus, it enables the attributor to maintain an illusion of objectivity while ensuring that only self-serving conclusions are reached.
A final word before concluding this discussion. For a number of years, there has been a debate over the mechanisms that underlie self-serving causal judgments (for reviews of this debate, see Brown & Rogers, 1990; Markus & Zajonc, 1985; Tetlock & Levi, 1982). Proponents of the motivational view (e.g., Pyszczynski & Greenberg, 1987; Zuckerman, 1979) assert that self-serving attributions and excuses are driven by a desire to perceive the self as good and efficacious. That is the motivational perspective assumes that seli-serving attributions are reached in order to defend, maintain, and promote high self-esteem. More cognitively-oriented theorists (Miller & Ross, 1975; Nisbett & Ross, 1980) have argued that recourse to motivational factors is unnecessary. These theorists argue that the self-serving attributional pattern can be accounted for entirely in terms of logical information-processing assumptions. After all, if one believes that one has the ability to succeed but fails, it is not entirely illogical to assume that failure must have been due to. some factor other than one's ability.
In the present view, research attempting to distinguish these two positions has clearly supported the motivational perspective (e.g., Brown & Rogers, 1990; Fries & Frey, 1980; Gollwitzer, Earle, & Stephan, 1982). But whatever the ultimate source of this bias turns out to be, the critical point for the present discussion is that the tendency to deny that failure implicates low ability enables individuals to maintain the belief that they are competent even when negative feedback -is encountered. In other words, the cognitive consequences of self-serving attributions are absolutely clear, even if their antecedents remain the source of some debate.
Strategies for Minimizing the Impact of Negative Feedback
Even the most exhaustive efforts to steer clear of negative feedback and avoid accepting responsibility for negative outcomes will not completely insulate individuals from unfavorable feedback. For some, years of struggling with even the most basic of household repairs provide compelling evidence that they lack mechanical ability. To simply deny the implications of such facts would be to experience repeated failure in that domain.
Under such conditions, individuals are apt to develop "acknowledged pockets of incompetence" (Taylor & Brown, 1988). That is, they readily admit to possessing the limitation in question, to the point where they may even exaggerate the extent of their deficiency. At the same time, however, most are able to do so in a manner that ensures that the damage to their overall sense of self-worth will be minimal. They accomplish this in part by calling on a host of reserve strategies that serve to maintain self-esteem. Five of these "damage-control" strategies are (a) selective importance, (b) selective consensus, (c) downward social comparison, (d) basking in reflected glory, and (e) compensatory self-enhancement.
Selective Importance
One way individuals can minimize the impact of an acknowledged limitation is by trivializing its importance. In support of this conjecture, substantial positive correlations have been found between an individual's perceived standing on an attribute and the perceived importance of that attribute (e.g., Rosenberg, 1979). Those who believe that they are intellectually gifted but inept in social situations tend to believe that intellectual ability is of greater importance than sociability. The reverse holds true for those who believe they perform better in social situations than in intellectual settings. To some extent, the tendency to see one's positive qualities as more important than one's negative qualities is quite logical. The individual who makes her living as an athlete may reasonably believe that coordination and strength are more important than are analytical and reasoning ability.
However, recent experimental research has documented the illusory nature of these beliefs. Gallagher and Brown (1989) first gave subjects information about a novel ability that purportedly measured flexibility in mental transformations. Based on their alleged performance on a test, some subjects were told that they had scored high in the ability and others were told that they had scored low in the ability. Consistent with the notion that individuals belittle the importance of things that they are deficient at, those given failure feedback subsequently judged the ability to be less important than did those who received success feedback (see also Johnson et al., 1983). By derogating the importance of the qualities that they believe they lack, individuals are able to accept a limitation and still ensure that its negative impact on the self is minimal.
Selective Consensus
A closely related strategy to the one documented above is a tendency to exaggerate the percentage of individuals who share one's limitation. Campbell (1986) found that although subjects underestimated consensus for their perceived abilities (e.g., few people can solve crossword puzzles as quickly as I can), they overestimated consensus for their perceived deficiencies (e.g., many people also have trouble with math). Viewing one's shortcomings as common allows the negative impact of an accepted liability to be softened (see also Snyder et al., 1983; Marks, 1984).
Downward Social Comparison
Another tactic for dealing with negative feedback involves downward social comparison (Brickman & Bulman, 1977; Wills, 1981). This strategy involves comparing oneself with those who are relatively disadvantaged on some dimension. A student who receives a "D" in a course can console himself by comparing with those * who failed the class. By focusing on those who are worse off than the self, one's own situation looks good in comparison. Recent research suggests that downward social comparison may be a particularly prevalent and adaptive strategy for coping with victimizing events. In a study of breast cancer patients, Wood, Taylor, and Lichtman (1985) found that the majority of women compared their condition with worse off others. Moreover, those who did so scored higher on measures of psychological adjustment. These findings suggest that downward social comparisons may be a ubiquitous and powerful strategy for maintaining self-worth following negative events.
Basking in Reflected Glory
Another strategy for dealing with negative feedback is in a sense the converse of downward comparison. Instead of comparing with those who are relatively disadvantaged on some dimension, one can attempt to augment self-worth after failure by emphasizing one's association with those who are relatively advantaged on some dimension. Cialdini and his associates have termed such a strategy basking in reflected glory (Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976; Cialdini & Richardson, 1980). In their research they found that students were more likely to use the pronoun we when discussing a football game their university team had won than when talking about a game their team had lost. Moreover, this tendency was most apparent after subjects had first experienced a personal failure. Emphasizing one's association with successful others appears to represent an additional method for minimizing the damage that failure causes to overall feelings of self-worth (see Tesser, 1988, for related research).
Compensatory Self-Enhancement
A final strategy that may be used to offset negative feedback regarding one aspect of the self is to exaggerate one's worth in other aspects of the self (Baumeister & Jones, 1978). An individual who has recently been rebuffed by his lover, for example, may try to counter this blow to self-worth by overemphasizing his ability in the athletic domain. In a test whether individuals engage in such compensatory self-enhancement tactics, Brown, Smart, and Gallagher (1990) first gave subjects positive or negative feedback on an alleged test of their intellectual ability. Subsequently, in an ostensibly unrelated experiment, subjects rated the degree to which a series of trait adjectives described them. Half of theattributes referred to qualities relevant to the achievement domain (e.g., smart, competent, etc.), the other half were interpersonal in nature (e.g., sincere, loyal, etc.). As predicted, although subjects' ratings of their achievement-related abilities were less favorable after failure than after success, their ratings of their interpersonal attributes were more favorable after failure than after success. Thus, subjects appeared to compensate for failure in the achievement domain by exaggerating their worth in the affiliative domain (see Steele, 1988, for related research).
Summary
A host of mechanisms that allow individuals to preserve their illusory self-conceptions have been explored in this section. Some of these strategies are designed to ensure that individuals receive predominantly positive feedback in their lives; others are designed to minimize the degree to which negative feedback implicates central aspects of the self; others represent attempts to preserve and restore self-esteem in the event that negative selfrelevant feedback is encountered and acknowledged. Together, they furnish individuals with an impressive array of weapons for staving off attacks on self-worth. (This discussion is not meant to be exhaustive. Other strategies for promoting self-worth related to those discussed above have been documented by Lewicki (1983, 1984), Wicklund and Gollwitzer (1982), and Conway and Ross (1984), among others.)
RELATED ISSUES
The preceding sections have documented the prevalence of self-enhancing illusions, their association with psychological adjustment and performance in achievement -related situations, and the means by which they are maintained. In the next section of this paper, consideration is given to some related issues.
Illusions and Mechanisms of Defense
As noted elsewhere (Snyder, 1985, 1988), illusions and the strategies that give rise to them are reminiscent of the psychoanalytic mechanisms of defense. Most notably, they both serve to protect an individual's sense of self-worth (cf. Hilgard, 1949). At the same time, however, there exist some important differences between these constructs. The first concerns their motivational bases. As conceived by Freud (1915/1957), defense mechanisms were presumed to be motivated solely by a desire to avoid pain and ward. off anxiety from internal threats. Later on, more ego-oriented theorists expanded Freud's idea -by postulating that defense mechanisms also served to protect the individual from threats that originated in the external world (see Ansbacher & Ansbacher, 1967). Despite this extension, defense mechanisms were still presumed to be invoked only in response to a perceived threat. Hence, the name defense. In contrast, self-enhancing illusions appear to be in service of both esteem enhancement and esteem protection. In other words, self-enhancing illusions may often be offensive, rather than solely defensive, in nature (see Sackeim, 1983, for an elaboration of this point).
A second important difference between defense mechanisms and self-enhancing illusions concerns
the role of the unconscious. According to psychoanalytic theory, defense mechanisms must occur outside of conscious awareness in order to be effective. This does not appear to be so for illusions. That is, in order to "work," illusions need not necessarily be restricted from awareness. For example, an individual may be fully cognizant of the fact that he will feel better about his tennis playing ability if he habitually plays with people whom he is likely to beat. The effectiveness of the illusion is not necessarily vitiated by an awareness of these facts.
In a related vein, self-enhancing illusions do not necessarily involve self-deception. According to most treatments of this issue, individuals are engaging in self-deception when they simultaneously hold two contradictory self-relevant beliefs: a positive belief that is fully conscious and a conflicting negative belief that is at least partly unconscious (Sackeim, 1983, 1988). Evidence of self-deception thus requires that at some less conscious level, individuals know that their beliefs are fanciful.
This does not appear to be the case with self-enhancing illusions. As indicated earlier, most illusions concern an individual's interpretation or construction of reality. Because reality can be construed in multiple ways, individuals are able to harbor most self-enhancing illusions without "resorting" to self-deception. Social attributes, for instance, are so amorphous that individuals can define them in nearly any way they choose. To see oneself as extremely sincere, for example, does not require self-deception (i.e., that at some level one doubts one's sincerity); it only requires an idiosyncratic definition of sincerity that emphasizes one's own perceived virtues (Dunning et al., 1989). Moreover, this is true even for attributes that are not social in nature. Some people (usually those who are agile and quick) define athletic ability in terms of balance and speed; others (who are usually brawny and thick) define athletic ability in terms of power and strength. Who's to say who's right? Hence, if both think that they are more athletic than the other, there is no reason to assume that any sort of deception is occurring.
Another related difference between defense mechanisms and illusions concerns the amount of distortion. Defense mechanisms may sometimes entail significant misrepresentations of physical reality. In extreme cases of repression, for example, an individual may entirely block out a threatening experience because it is too painful to acknowledge. Illusions, in contrast, typically do not involve major distortions of reality. Rather, they entail important variations in how reality is interpreted or represented. A woman who is suffering through a divorce, for example, may comfort herself in a number of ways. She could belittle the importance of marriage as a life-style, remind herself of the number of other couples who have separated due to marital difficulties, compare herself with those who are helplessly trapped in unhappy marriages, and so on. All of these strategies are likely to make her feel better, but none involve a major distortion of reality.
A final distinction worth noting between defense mechanisms and illusions concerns the degree to which each is responsive to environmental feedback. Whereas defense mechanisms tend to be held more rigidly as threats to the self increase, illusions are relatively responsive to threat (Taylor, 1988). For instance, earlier it was noted that rather than denying the implications of pervasive negative feedback in some domain of life, individuals accept the limitation but compensate by exaggerating their worth in other respects of life (Brown et al., 1990; Taylor, Collins, Skokan, & Aspinwall, 1989). These findings suggest that illusions are more flexible and accommodating to the environment than are the defense mechanisms.
Potential Risks
A complete account of illusions must include a consideration of their potential liabilities. To be sure, self-enhancing illusions involve some inherent risks. Overly positive views of the self may lead individuals to pursue goals that are beyond their capabilities; an exaggerated sense of control may lead people to blame themselves for negative outcomes that were truly not within their control; excessive optimism may lead individuals to underestimate their susceptibility to some diseases, thereby limiting the likelihood that they will take appropriate preventive measures. Thus, associated with each of the illusions considered above is an attendant set of risks.
The extent to which these potential risks mitigate the effectiveness of illusions, however, is unclear. Because they are positively related to psychological adjustment and performance outcomes, whatever limitations inherent in illusions appear to be outweighed by the benefits they provide. One possible explanation for this fact has been offered by Taylor and Brown (1988). These theorists speculated that the negative aspects of illusions may cancel out one another. For instance, although unrealistic optimism may diminish the chances that individuals will initiate proper health behaviors, a strong sense of mastery and the belief that one can, control events may lead them to assiduously maintain proper health habits once they are begun (cf. Bandura, 1988). In this manner, illusions may operate in a sort of "checks and balances" fashion that ensures that their potential liabilities do not exceed their advantages. However, there is currently a paucity of research on this matter, and investigations regarding the potential liabilities of illusions represent an important topic for future research.
Implications for Psychotherapy
Assuming that the limitations of illusions are minimal, or at least manageable, the present analysis would seemingly have important implications for psychotherapy. As indicated earlier, many, though not all, approaches to psychotherapy stress that therapeutic change requires the acquisition of accurate self-knowledge. For instance, therapies developed by humanistic psychologists (e.g., Perls, 1973; Rogers, 1951) encourage individuals to "get in touch with their true selves"; many cognitive-behavioral therapies (e.g., Beck, 1967) urge individuals to carefully monitor their self-relevant thoughts, in the hope that distorted patterns of thinking can be corrected (i.e., made to be more realistic); finally, even psychoanalysis is based on the notion that treatment should involve a search for truth. Thus, although different in many ways, the tie binding these approaches is that accurate self-knowledge will help ameliorate individuals' distress.
The research reviewed in this paper suggests that this assumption may be unfounded. As we have seen, at least with respect to the most common psychological disorder (i.e., depression), accurate self-understanding appears to be negatively related to adjustment. This admits the possibility that rather than encouraging individuals to gain unbiased self-knowledge, greater therapeutic progress might be achieved by encouraging individuals to acquire self-enhancing illusions (Alloy & Abramson, 1988).
How might such an intervention proceed? A useful starting point might be to train clients in the strategies normal individuals use for enhancing self-worth. Interestingly, these strategies tend to be less evident among individuals who are psychologically distressed (i.e., depressed, anxious, or low in self-esteem). For example, compared with normals, distressed individuals are less likely to (a) interpret ambiguous feedback to be positive feedback (Dykman & Volpicelli, 1983); (b) selectively attend to and remember positive self-relevant feedback over negative self-relevant feedback (Nelson & Craighead, 1977; Johnson et al., 1983); (c) attribute positive outcomes to the self but deny responsibility for negative outcomes (Campbell & Fairey, 1985; Kuiper, 1978); (d) judge their positive attributes to be more important than their deficiencies (Wenzlaff & Grozier, 1988); (e) view their assets as unique but their liabilities as commonly shared by others (Campbell, 1986); and (f) show compensatory self-enhancement following failure (Baumeister, 1982; Brown et al., 1990). In short, like the illusions they give rise to, strategies for promoting self-worth are relatively absent among distressed individuals. By exposing clients to these strategies and encouraging their judicious use, therapists might be able to help their clients experience improvements in symptomatology (cf. Sackeim, 1983; Shrauger, 1982).
Of course, before such interventions can be implemented on anything more than an experimental basis, more evidence pertaining to the causal role of illusions in promoting mental health needs to be amassed. To date, most of the research regarding illusions and adjustment is correlational. However, some experimental evidence linking illusions to emotional well-being does exist. For instance, McFarland and Ross (1982) found that leading subjects to make self-serving attributions resulted in greater positive affect after success and lower negative affect after failure. There is also evidence that encouraging subjects to focus on their likely success at an upcoming task actually improves task performance (Campbell & Fairey, 1985). These are just a few of the ways that self-enhancing illusions might promote emotional health and psychological well-being (for additional evidence, see Snyder & Higgins, 1988; Taylor & Brown, 1988). If future research continues to document causal relations between illusions and adjustment, treatment regimens based on encouraging illusory thinking would seem to represent a viable therapeutic intervention.
NEED FOR ACCURATE
SELF-KNOWLEDGE
To conclude, it is appropriate to reconsider the question that was posed earlier in this paper: what is the proper relation between an individual's picture of the self and the "actual" self? Although theorists from a variety of disciplines have asserted that individuals ought to, and need to, possess accurate and realistic self -perceptions, the present review suggests that most individuals possess views of the self that are more favorable than can realistically be justified, and that this tendency is associated with, and may contribute to, psychological adjustment and superior functioning. Thus, setting aside issues of morality, the answer to this question seems to be that the proper relation is one in which one's picture of the self is brighter, more colorful, and more beautiful than the "actual" self.
An appropriate analogy can perhaps be drawn between self-enhancing illusions and the new magnetic levitation vehicles that are currently being developed for use in Japan, France, and West Germany. These passenger trains are capable of achieving speeds of up to 300 miles per hour by riding an electromagnetic current that raises them just slightly above the rails. The trick is in keeping the train just the right distance off the ground; rising too high causes the train to gyrate and crash; riding too close to the ground causes the train to grind to a halt.
In a similar vein, self-enhancing illusions are probably most effective when they are only slightly more positive than can realistically be justified (Baumeister, 1989). Being too grandiose in one's thinking can have serious consequences, as the destructive delusions of grandeur that can accompany manic-depressive illness illustrate. But being too modest in one's self-appraisals can also be debilitating, as research on depressive realism attests (Alloy & Abramson, 1988). Thus, much like the new passenger trains under development, individuals' self-appraisals appear to be most effective when they rise slightly above the ground.