TO BE USED IN GPS 520 ONLY

This article is for use only of students enrolled in GPS 520: Advanced Social Psychology for purposes associated with this course and may not be retained or further disseminated.

Excerpts taken from Brown, J. D. (1998). The self (pp. 190-228). Boston, MA: McGraw-Hill

Self-esteem is the panacea of modern life. It is seen as the key to financial success, health, and personal fulfillment, and it is regarded as the antidote to underachievement, crime, and drug abuse (Branden, 1994; Mecca, Smelser, & Vasconcellos, 1989). Self-esteem is also popular in academic circles. In the fields of personality and social psychology, it has been implicated in models of conformity (Brockner, 1984), attraction (Hatfield, 1965), persuasion (Rhodes & Wood, 1992), cognitive dissonance (Steele, Spencer, & Lynch, 1993), subjective well-being (Diener & Diener, 1995), and social comparison processes (Aspinwall & Taylor, 1993; Gibbons & Gerrard, 1991; Wood, Giordano-Beecli, Taylor, Michela, & Gaus, 1994), just to name a few.

The widespread appeal of self-esteem attests to its importance, but this popularity has had an undesirable consequence. Self-esteem is currently spread so thin that it is difficult to know just what it is. It is used as a predictor variable (some researchers study whether high self-esteem people think, feel, and behave differently than low self-esteem people), an outcome variable (some researchers study how various experiences affect the way people feel about themselves), and a mediating variable (the need for high self-esteem is presumed to motivate a wide variety of psychological processes). In short, self-esteem has become a protean concept—so capable of changing form that its value is in risk of being undermined.

In this chapter we will critically review the nature, origins, and functions of self-esteem. The chapter begins by considering the nature of self-esteem. Here we ask: What do we mean by the term “self-esteem,” and what characteristics are associated with high self-esteem and with low self-esteem? Next we examine the genesis of self-esteem. Our concern here will be with understanding what experiences give rise to high self-esteem and to low self-esteem. We will then consider when self-esteem is important. Here we will ask: What difference does it make whether a person has high self-esteem or low self-esteem? Finally, we will consider some unresolved controversies in the field.

A central issue guiding our discussion will involve the nature of self-esteem itself. Some psychologists (myself included) adopt an affective approach to understanding the nature of self-esteem. This approach asserts that self-esteem is a feeling of affection for oneself that develops largely through visceral or irrational processes (irrational in this case means “not based in logic”). Other psychologists take a more cognitive view. They believe that self-esteem is a judgment people make about themselves. This judgment is largely based on an assessment of one’s various abilities and attributes. Whereas the former approach likens self-esteem to feelings of love (which typically are not rational or logical), the latter likens self-esteem to a decision people make about their worth and value as a person. Although not all theories fall neatly into one camp or the other, keeping this affect—cognition distinction in mind will help you organize the material that follows.

One more word before we begin. Self-esteem has been implicated in many psychological phenomena of a clinical nature (e.g., anxiety, depression, eating disorders, and substance abuse) (for a review, see Robson, 1988). Issues of this nature will he discussed in Chapter 9. The current chapter will focus on self-esteem within normal (i.e., nonclinical) populations.

Self-esteem is part of everyday language, and at an intuitive level, everyone seems to know what self-esteem “is.” It may surprise you then to know that there is less than perfect agreement within the psychological literature. Part of the problem is that the term is used in three different ways.

Most often, the term “self-esteem” is used to refer to a personality variable that captures the way people generally feel about themselves. Researchers call this form of self-esteem global self-esteem or trait self-esteem, as it is relatively enduring, both across time and situations. In this book, I have used the term self-esteem, (without any qualifiers) when referring to this variable.

Attempts to define self-esteem have ranged from an emphasis on primitive libidinal impulses (Kernberg, 1975), to the perception that one is a valuable member of a meaningful universe (Solomon, Greenberg, & Pyszczynski, l991). I take a decidedly less exotic approach and define self-esteem in terms of feelings of affection for oneself (Brown, 1993; Brown & Dutton, 1995b). Within normal populations, high self-esteem is characterized by a general fondness or love for oneself; low self-esteem is characterized by mildly positive or ambivalent feelings toward oneself. In extreme cases, low self-esteem people hate themselves, but this kind of self-loathing occurs in clinical populations, not in normal populations (Baumeister, Tice, & Hutton, 1989).

The term self-esteem is also used to refer to the way people evaluate their various abilities and attributes. For example, a person who doubts his ability in school is sometimes said to have low academic self-esteem, and a person who thinks she is popular and well liked is said to have high social self-esteem. In a similar vein, people speak of having high self-esteem at work or low self-esteem in sports. The terms self-confidence and self-efficacy have also been used to refer to these beliefs, and many people equate self-confidence with self-esteem. I prefer to call these beliefs self-evaluations or self-appraisals, as they refer to the way people evaluate or appraise their abilities and personality characteristics.

Self-esteem and self-evaluations are related—people with high self-esteem think they have many more positive qualities than do people with low self-esteem—but they are not the same thing. A person who lacks confidence in school might still like himself a lot. Conversely, a person who thinks she is attractive and popular might not feel good about herself at all. Unfortunately, psychologists don’t always make this distinction, often using the terms self-esteem and self-evaluations interchangeably.

The causal association between self-esteem and self-evaluations is also unclear. Cognitive models of self-esteem assume a bottom-up process (e.g., Harter, 1986; Marsh, 1990; Pelham & Swann, 1989). They assume that positive evaluations of self in particular domains give rise to high self-esteem. I call this a bottom-up process because it assumes that global self-esteem is built up from these more specific evaluations. Affective models of self-esteem assume a top-down process (Brown, 1993; Brown, Dutton, & Cook, 1997). These models assume that the causal arrow goes from global self-esteem to specific self-evaluations: Liking oneself in a general way leads people to believe they have many positive qualities. Later in this chapter we will examine support for these claims.

Finally, the term self-esteem is used to refer to rather momentary emotional states, particularly those that arise from a positive or negative outcome. This is what people mean when they speak of experiences that bolster their self-esteem or threaten their self-esteem. For example, a person might say her self-esteem was sky-high after getting a big promotion, or a person might say his self-esteem was really low after a divorce. Following William James (1890), we will refer to these emotions as self-feelings or as feelings of self-worth. Feeling proud or pleased with ourselves (on the positive side), or humiliated and ashamed of ourselves (on the negative side) are examples of what we mean by feelings of self-worth.

Because they involve feelings toward oneself, some researchers (e.g., Butler, Hokanson, & Flynn, 1994; Leary, Tambor, Terdal, & Downs, 1995) use the term state self-esteem to refer to the emotions we are calling feelings of self-worth, and trait self-esteem to refer to the way people generally feel about themselves. These terms connote an equivalency between the two phenomena, implying that the essential difference is simply that global self-esteem is persistent, while feelings of self-worth are temporary.

The trait—state assumption has important consequences. First, it suggests that feeling proud of oneself is akin to having high self-esteem and that feeling ashamed of oneself is akin to having low self-esteem. This, in turn, leads investigators to assume that an analogue of high self-esteem or low self-esteem can he created by temporarily leading people to feel good or bad about themselves (e.g., Greenberg et al., 1992; Heatherton & Polivy, 1991; Leary et al., 1995). This is typically accomplished by giving people positive or negative self-relevant feedback (e.g., telling people they are high or low in some ability). Other researchers disagree with this approach, arguing that these manipulations do not provide a suitable analogue of high self-esteem or low self-esteem (Brown & Dutton, 1995b; Wells & Marwell, 1976).

One more point about feelings of self-worth. Several times in this book we have spoken of a basic human need to feel good about ourselves. Within psychology, this is called the self-enhancement motive. This term refers to the fact that people are motivated to have high feelings of self-worth. People want to feel proud of themselves rather than ashamed of themselves. They strive to maximize and protect their feelings of self-worth. The way people go about trying to meet this need differs across time, cultures, and subcultures, but the need is universal. The conclusion was perhaps best stated by the Pulitzer Prize-winning anthropologist Ernest Becker, who wrote:

The fundamental datum for our science is a fact that at first seems banal, or irrelevant: it is the fact that—as far as we can tell—all organisms like to ‘feel good” about themselves Thus in the most brief and direct manner, we have a law of human development.... (Becker, 1968, p. 328)

Interestingly, there is no consensus on why people are motivated to have positive feelings of self-worth. Some believe these feelings are intrinsically satisfying; as James (1890) put it, “direct and elementary endowments of our nature” (1890, p. 306). Others (Gergen, 1971; Kaplan, 1975) believe that positive feelings of self-worth are preferred simply because they have come to be associated with positive outcomes, such as praise from others or success. Still others believe that feelings of self-worth are desired because they imbue life with meaning and make one’s inevitable death more tolerable (Greenberg et al., 1992). Whatever the source of this need may be, a desire to promote, maintain, and protect positive feelings of self-worth has been assumed to motivate a wide range of human behavior. This includes behavior in achievement settings (Covington & Berry, 1976), social settings (Tesser, 1988), and health settings (Ditto & Lopez, 1992).

Now that we have some idea of what we mean by self-esteem, we can look at how it is measured. You probably know someone who you think has low self-esteem. Your intuitions are probably based on what the person says and the things the person does. Psychologists also rely on these cues to measure self-esteem (Demo, 1985).

_____________________________________________________________________________________________

Indicate your level of agreement with each of the following statements by circling one number on the rating scale that best describes the way you feel about yourself. Use the following scale as your guide.

Note: To determine your score, first reverse the scoring for the five negatively worded items (1, 3, 4, 5, and 8) as follows: 0 = 3, 1 = 2, 2 = 1, 3 = 0. Then, add up your scores across the 10 items. Your total score should fall between 0 and 30. Higher numbers indicate higher self-esteem.

The Rosenberg (1965) self-esteem scale is one of the most widely used instruments for measuring self-esteem in research settings. This scale, which is shown in Table 8.1, was developed to assess global self-esteem. It focuses on people’s general feelings toward themselves, without referring to any specific quality or attribute. Half of the items are worded in a positive direction (“On the whole, I am satisfied with myself”); the other half are worded in a negative direction (“All in all, I am inclined to feel that I am a failure”).

Table 8.2 presents another widely used measure of self-esteem, the Texas Social Behavior Inventory (Helmreich & Stapp, 1974). This scale is often used as a measure of global self-esteem, but it actually measures how comfortable and competent a person feels in social situations. Scores on this scale are related to scores on the Rosenberg self-esteem scale (r = .65 or so), but the two scales do not measure the same thing. A person can be uncomfortable in social situations and still like himself in general. Alternatively, a person can be relaxed and outgoing with others but not like herself in general. For this reason, the Rosenberg scale is the appropriate one to use for measuring global self-esteem.

_____________________________________________________________________________________________

Note: To determine your score, first reverse the scoring for the negatively worded items (numbers 1, 5, 6, 12, 13, and 15) as follows: 1 = 5, 2 = 4, 3= 3, 4 = 2, and 5 = 1). Then, add up your scores for the l6 items. Your total score should fall between 0 and 80. Higher numbers indicate higher self-esteem.

Many other self-report instruments are available (Blascovich & Tomaka, 1990). Based on theoretical work by Shavelson, Hubner, and Stanton (1976; see also, Byrne & Shavelson, 1996), Marsh (1990) developed an extensive measure to assess the way people evaluate themselves in various domains of life. Included are items pertaining to one’s (perceived) physical abilities, appearance, problem-solving abilities, social skills, peer relationships, opposite-sex relationships, and emotional stability. Harter (1986) has developed a similar scale for children, with subscales assessing (perceived) scholastic competence, athletic competence, social acceptance, physical appearance, and behavioral conduct. Scales of this type focus on the second meaning of self-esteem we discussed earlier: They assume that people have different self-esteem levels for different attributes, situations, and activities. Typically, these scales also include a separate subscale to measure global self-esteem.

Self-report measures of self-esteem are widely used and possess a high degree of theoretical and predictive validity (Rosenberg, 1979; Wells & Marwell, 1975). But they are not without problems. For example, Baumeister, Tice, and Hutton (1989) have argued that scores on self-report measures of self-esteem are compromised by self-presentational concerns. Rather than rating how they really feel about themselves, people may distort their responses to create a particular impression in the minds of other people. From this perspective, a high self-esteem score represents an assertive, interpersonal style, in which one is willing to present oneself to others in a highly positive manner, and a low self-esteem score reflects a modest interpersonal style, in which one is reluctant to present oneself in a highly positive matter (see also, Arkin, 1981; Hill, Weary, & Williams, 1986). These self-presentational patterns are assumed to be imperfectly correlated with how people privately feel about themselves.

Defensive processes also influence self-report measures of self-esteem (Weinberger, 1990; Westen 1990b). People who score high on self-report measures of self-esteem may be fooling themselves by defensively claiming to feel better about themselves than they really do. On the other hand, some forms of self-deception may actually be healthy and play an integral role in psychological adjustment (Paulhus & Reid, 1991; Sackeim, 1993; Taylor & Brown, 1988, 1994). We will look at this issue more closely in Chapter 10.

Greenwald and Banaji (1995) have suggested adapting methods from cognitive psychology to overcome these potential limitations of self-report measures. These indirect or implicit measures of self-esteem (e.g., response latencies, recognition thresholds) would be less transparent than self-report questionnaires, in that people would be unaware that their self-esteem was being assessed. Although such measures have yet to be developed, it is likely that they will be available in the years to come.

Having defined our terms and discussed ways to measure self-esteem, let’s look more closely at the nature of self-esteem. What aspects make up high self-esteem? And where does high self-esteem come from? Several different models have been developed to address these questions.

Affective models of self-esteem assume that self-esteem develops at an early age and is characterized by two types of feelings. One of these feelings (which we will call feelings of belonging) is rooted in social experiences; the other (which we will call feelings of mastery) is somewhat more personal in nature.

Belonging is the feeling that one is unconditionally loved and valued, not for any particular quality or reason but simply for who one is. A sense of belonging gives people a secure base in life. It gives them the feeling that no matter what happens, they are valued and respected. Some years ago, the American psychologist Carl Rogers highlighted this aspect of self-esteem when he discussed people’s need for unconditional positive regard (Rogers, 1951; Rogers & Dymond, 1954).

Feelings of belonging are a bit different than reflected appraisals. As discussed in Chapter 3, reflected appraisals represent our conscious perception of how we are viewed by others. If I think other people think I’m funny, then I think I’m witty. Feelings of belonging do not occur at a conscious level. They are more intuitive. Belonging is the feeling that we are loved and the security that feeling brings.

The second important aspect of self-esteem is a sense of mastery. Mastery involves the perception that one is having an impact on the world—not necessarily in any large-scale sense, but in one’s day-to-day life. Mastery is not the same as perceived competence, although some writers (e.g., Tafarodi & Swann, 1995) use the two terms interchangeably. We needn’t think we are an accomplished pianist or an “A” student to develop a sense of mastery. Rather, mastery is the feeling we get when we are immersed in an activity or are striving to overcome some obstacle (e.g., Brissett, 1972; Csikszentmihalyi, 1975; deCharms, 1968; Deci & Ryan, 1995; Erikson, 1963; Franks & Marolla, 1976; Gecas & Schwalbe, 1993; White, 1959).

One way to convey the difference between mastery and perceived competence is to consider a child who is making mud pies. The squishing, the feeling of the mud between the child’s fingers and the sheer joy that comes from that experience creates feelings of mastery. These feelings promote high self-esteem. But this is not the same as thinking one is a “good mud pie maker.” The squish is process oriented—it is joy in creating and manipulating; the evaluation is outcome oriented—it is a judgment of whether one is good at something. The affective model maintains that only the former is relevant to the genesis of self-esteem.

Affective models of self-esteem assume that feelings of belonging and mastery normally develop early in life. Erik Erikson’s model of psychosocial development (discussed in Chapter 4) provides a useful springboard for considering how these feelings arise. According to Erikson, the first developmental task infants face is establishing feelings of trust with their caregivers. These feelings of trust, which are thought to develop during the first year of life, correspond to the feelings of belonging we have said are integral to a sense of high self-esteem.

The next stage Erikson describes is the “autonomy versus shame and doubt” stage. This stage involves the development of feelings of mastery. Children develop feelings of mastery when they are encouraged to explore, create, and modify their world (e.g., to build things, to draw, or paint); they may fail to develop these feelings when their parents subvert, ridicule, or are overly critical of their efforts (Stipek, Recchia, & McClintic, 1992).

The caregiver-child relationship plays a key role in Erikson’s theory. This relationship also plays a central role in other theories of self-esteem development (e.g., Baumeister & Leary, 1995; Bowlby, 1969; Epstein, 1980; Sullivan, 1953). Bowlby’s (1969) attachment theory is particularly relevant to the present discussion. Bowlby was interested in understanding the basis and functions of attachment bonds. He noted that in humans, as well as in other animals, infants bond with their caregivers (particularly with their mothers). Why? What function do these mother—child bonds serve?

Bowlby surmised that the attachment relationship serves a paradoxical function. By becoming securely attached, the child feels safe enough to leave the mother and explore the environment. In this sense, Bowlby believed that a feeling of belonging (i.e., a secure attachment) facilitates a sense of mastery (willingness to explore the environment).

When individuals of any age are feeling secure they are likely to explore away from their attachment figure. When alarmed, anxious, tired, or unwell they feel an urge toward proximity. Thus, we see the typical pattern of interaction between child and parent known as exploration from a secure base. Provided the parent is known to be accessible and will be responsive when called upon, a healthy child feels secure enough to explore. (Bowlby. 1979, p. 3)

A series of studies using a procedure known as the strange situation (Ainsworth, Blehar, Waters, & Wall, 1978) has documented these effects. In this situation, a young child (typically around 14 months of age) is brought into a psychological laboratory with his or her mother. The room contains a number of interesting toys and objects that most children enjoy looking at and playing with. The extent to which the child initially explores the objects in the room is one variable of interest.

Another variable of interest is how the child reacts to separation from the mother. After being together for a few minutes, the mother unexpectedly leaves the child alone with a stranger. The child’s emotional reaction to the mother’s departure is noted. Several minutes later, the mother returns and the researcher notes the child’s emotional and behavioral reaction to the mother’s return. In this manner, the strange situation measures the extent to which a child uses the mother as a secure base from which to explore the environment and as a source of comfort in times of stress.

In studies using this procedure, three different attachment styles have been identified.

1. Approximately 60 percent of American infants are classified as being securely attached. Securely attached infants show a healthy balance between closeness to the mother and independence. During the first phase of the procedure, they readily separate to explore the environment. Although they may he distressed when their mother leaves, they are eager to see her when she returns and enjoy drawing her into their play and sharing their discoveries with her.

2. Approximately 15 percent of American infants are classified as anxious/ambivalent. These children have difficulty separating during the first phase of the procedure. They are unwilling or afraid to explore the environment. When their mother leaves, they become very distressed and upset. Although they are somewhat comforted when she returns, they cling to her and show other signs of insecure dependence (e.g., they continue whining).

3. Approximately 25 percent of American infants are classified as avoidant children. These children tend to avoid or ignore their mothers altogether. They appear to have little difficulty separating during the first phase of the procedure, and they outwardly exhibit few signs of distress when their mother leaves. Furthermore, they show little interest in her when she returns, preferring instead to play alone rather than to interact with her. Importantly, the indifference these infants display toward their mothers is contradicted by an inner sense of anxiety and distress. Rather than being secure and independent, avoidant children are evading intimacy and closeness with their mothers.

The roots of self-esteem would seem to lie within these different attachment styles. The avoidant infants may develop feelings of mastery (because they willingly explore the environment), but they lack a sense of belonging. They do not exhibit a strong emotional bond to their mother. The anxious/insecure infants may display a sense of belonging, but they are unlikely to develop feelings of mastery. They are easily distressed and are unwilling to meet the world head on. Only the securely attached children exhibit both a strong sense of belonging and a strong sense of mastery. It is these children, then, who are most apt to develop high self-esteem.

Research supports this conjecture. Different attachment styles in infancy predict self-esteem in preschool and kindergarten (Cassidy, 1990; Sroufe, 1983), with securely attached children showing the highest self-esteem. Similar patterns have been found with adolescents and young adults (Bartholomew & Horowitz, 1991; Brennan & Morris, 1997; Collins & Read, 1990; Feeney & Noller, 1990; Griffin & Bartholomew, 1994).

Bowlby (1973) invokes the concept of an “internal working model” to explain why the early attachment relationship has an enduring effect. As children mature, they develop a cognitive representation or working model of the attachment relationship. A child who develops a secure attachment relationship comes to believe she is essentially good and worthy of love; a child who develops an insecure attachment comes to believe she is bad and unworthy of love. These beliefs generalize to other people and situations and form the basis for the development of self-esteem.

An unwanted child is likely not only to feel unwanted by his parents but to believe that he is essentially unwantable, namely unwanted by anyone. Conversely, a much-loved child may grow up to be not only confident of his parents’ affection but confident that everyone else will find him lovable too. Though logically indefensible, these crude overgeneralizations are nonetheless the rule. Once adapted, moreover, and woven into the fabric of working models, they are apt henceforward never to be seriously questioned. (Bowlby, 1973, pp. 204—205)

Affective approaches to understanding self-esteem make the following points: (1) unconditional feelings of belonging and a sense of mastery comprise the essence of high self-esteem, and (2) these feelings typically develop early in life, largely as a result of parent—child interactions. This emphasis on early childhood experiences does not mean that self-esteem can never change. It simply means that early experiences lay the foundation for high self-esteem or low self-esteem. Later experiences in life may also affect self-esteem, although none is apt to be as important as the parent—child relationship.

One reason that latter experiences are less consequential is that they are always viewed through the prism or schema that is established earlier. Once high or low self-esteem develops, it guides the way we view ourselves, other people, and the experiences and events we confront. Often, this guiding process occurs at an automatic or preconscious level (Epstein, 1990), making it difficult to detect and even harder to correct. For this reason, self-esteem tends to persist.

Cognitive models offer a different perspective on the nature and origins of self-esteem. They view self-esteem as a more or less conscious decision people make regarding their worth as a person. If you think you possess many socially desirable qualities, then you will have high self-esteem. In terms of the three meanings of self-esteem we discussed earlier, cognitive models emphasize that how we evaluate ourselves in various domains determines our overall level of self-esteem.

The simplest of these models assumes that self-esteem is the aggregate of the way people evaluate their specific qualities and attributes. The top portion of Table 8.3 illustrates this add-em-up approach. Here we have asked two (imaginary) people to indicate how attractive, intelligent, well liked, and athletic they think they are using seven-point scales (e.g., I = not at all attractive; 7 = very attractive). Person A thinks he is quite attractive, not terribly intelligent, reasonably well liked, and very athletic; Person B thinks he is not terribly attractive, very intelligent, moderately well liked, and not very athletic.

According to the add-em-up approach, we would simply add up these various scores to determine the person’s overall level of self-esteem. In this example, we would predict that Person A has higher self-esteem than Person B. This is because Person A evaluates himself more positively than does Person B.

_____________________________________________________________________________________________

The add-em-up model assumes that global self-esteem represents the sum of the way people evaluate their more specific qualities. To test this approach, we would simply add up the four self-evaluation scores to determine each person’s self-esteem score. Using this approach, we would predict that Person A has higher self-esteem than Person B.

_____________________________________________________________________________________________

The weight-em-by-importance model assumes that self-esteem depends not only on how you evaluate yourself in specific domains but also on how important you think it is to be good in those domains. To test the model, we have each person rank the four attributes in terms of their personal importance (1 = least important; 4 = most important). We then multiply each self-evaluation score by its corresponding importance rating (in parentheses) and add the products. Using this approach, we would predict that Person B has higher self-esteem than Person A. This is because Person B values what he thinks he is good at more than Person A.

_____________________________________________________________________________________________Self-Ideal Model

The self-ideal model assumes that self-esteem depends on the difference between who we think we are now and who we would ideally like to be. To test the model, we have each person indicate how attractive, intelligent, well liked, and athletic they would like to be (1 = not at all; 7 very). We then subtract these ideal self-ratings (in parentheses) from their corresponding self-evaluation score and then sum the differences. Using this approach, we would predict that Person B has higher self-esteem than Person A.

_____________________________________________________________________________________________

Note: For each example, two (hypothetical) people have indicated how attractive, intelligent, well liked, and athletic they think they are (1 not at all; 7 = very).

One problem with this method (which you may have already identified) is that it ignores the fact that different things are important to different people. if athletic ability is unimportant to Person A, and intellectual ability is extremely important to Person B, then Person B may feel better about himself than Person A.

The idea that self-esteem depends on what you think about yourself in domains of high personal importance is reminiscent of James’s (1890) claim that “self-esteem = success/pretensions.” In Chapter 2 we noted that James uses the word pretensions in two ways, Sometimes it refers to what we value in life or to what we think is important. Here James is saying that outcomes in domains of high personal importance have a greater effect on self-esteem than do outcomes in domains of low personal importance. James also uses pretensions to refer to a person’s level of aspiration. In this case, he is saying that we feel good about ourselves when our outcomes exceed our personal standards and bad about ourselves when our outcomes fall short of our personal standards.

Of the two meanings, most contemporary psychologists have focused on the one that emphasizes the importance of different attributes for self-esteem. Morris Rosenberg stated the case for the “importance of different attributes” as follows:

Ordinarily, we assume that if someone respects himself in certain particulars, then he respects himself in general. If he thinks he is smart, attractive, likable, moral, interesting, and so on, then he thinks well of himself in general. Yet it should be apparent that. . . a person’s global self-esteem is based not solely on an assessment of his constituent qualities but on an assessment of the qualities that count. . . . The differential importance of self-concept components is thus critically significant for self-esteem. (Rosenberg, 1979, p. 18)

The middle portion of Table 8.3 illustrates one way to test this weight-em-by-importance model. For this example, we have asked the two people to rank the four attributes in terms of their importance (1 = least important; 4 = most important). We have then multiplied their attribute ratings by their importance ratings (in parentheses), and then added up the products to form a weighted self-esteem score. Now we would predict that Person B feels better about himself than does Person A. This is because Person B values what he thinks he is good at more than Person A values what he thinks he is good at (Pelham, 1995; Pelham & Swann, 1989).

Despite its intuitive appeal, research has not found strong support for the weight-em-by-importance model (Marsh, 1993b, 1995; Pelham, 1995). Simply adding up the person’s self-evaluations and ignoring importance often provides as good an indication (if not better) of the person’s level of self-esteem. This may be because people tend to believe all of these attributes are important, so that the importance rating does not add much information. Another possibility is that it is not the individual’s own importance rating that is critical, but how important the attribute is to society in general (Hoge & McCarthy, 1984; Marsh, 1993b, 1995). This possibility assumes that individuals are not entirely free to decide what is important and what is not.

A final approach to understanding self-esteem looks at the discrepancy between the way people view their specific qualities and their ideal of who they should be in that domain (Higgins, Klein, & Strauman, 1985; Homey, 1945; Rogers, 1951, 1954). This approach also derives from James’s (1890) formula, but here we are treating pretensions as level of aspiration—what kind of person do you want to be, think you should be, or ought to be—rather than as values. The more our current self-image matches these idealized self-images, the higher is our self-esteem.

One way to test this model is to have people indicate how they would like to see themselves in various domains (e.g., “How intelligent would you like to be?”). We then subtract these ideal self-ratings from the person’s current self-evaluation. The bottom portion of Table 8.3 presents a hypothetical example. Person A is a perfectionist. He needs to be great at everything. Consequently, although he evaluates himself highly, he falls short of his ideals and we would predict that he has low self-esteem. Person B doesn’t evaluate himself as highly, but he doesn’t think he has to be “great” at everything either. So, we predict that he has high self-esteem.

Empirical tests of this model have found support for the claim that high self-esteem is associated with small “self—ideal self” discrepancies (Higgins et al., 1985; Ogilvie, 1987; Rogers & Dymond, 1954). Unfortunately methodological problems associated with the use of difference scores cloud the interpretation of these findings (Wylie, 1979).

Having examined three models that relate self-esteem to self-evaluations, let’s look more closely at the relation between these two variables. Table 8.4 presents data relevant to this issue. The data come from a study I conducted at the University of Washington. In this study, college students evaluated themselves and most other people on a number of abilities and personality attributes. I then examined these ratings as a function of the students’ self-esteem level, as measured by the Rosenberg (1965) self-esteem scale.

The table shows a number of interesting effects. The top half of the table shows that self-esteem is strongly related to people’s self-evaluations. Using a traditional .05 level of significance, high self-esteem students rated themselves more favorably than did low self-esteem students on 11 of the 14 positively valued traits and on 7 of the 12 negatively valued attributes. This means that self-esteem effects are quite broad. In comparison with low self-esteem people, high self-esteem people think they are better at a great many socially valued traits.

The bottom half of Table 8.4 shows that both self-esteem groups describe themselves in more positive (or less negative) terms than they describe most other people, and that this tendency is especially pronounced among high self-esteem students. This finding is of interest because it has been suggested that high self-esteem people show a comparable high regard for others, whereas low self-esteem people derogate others in an attempt to compensate for feelings of inadequacy (e.g., Epstein & Feist, 1988; Fromm, 1963; Rogers, 1951). Although this may occur when people evaluate members of their extended self (e.g., close friends and loved ones), it does not occur when people evaluate “most other people.” The tendency to see oneself as “better than average’ is far more characteristic of high self-esteem people than of low self-esteem people (Brown, 1986).

A final thing to note is that these low self-esteem students do not really describe themselves in negative terms. Their self-evaluations are less positive than are those of the high self-esteem students, but they are not negative in any absolute sense. In fact, they are quite positive. This is especially true when we compare their self-evaluations with their evaluations of most other people. Low self-esteem students appraised themselves more positively than most other people on 8 of the 14 positive attributes and less negatively than most other people on all 12 of the negative attributes. In many cases, these differences are substantial. For example, low self-esteem students regarded themselves as much more compassionate, kind, and loyal than are most other people, and as much less inconsiderate, unattractive, and unintelligent.

What we have, then, is a tendency for high self-esteem people to regard themselves as good at just about everything and a tendency for low self-esteem people to see themselves in generally, but not excessively, positive terms (Baumeister et al., 1989; Brown, 1986, 1993). Keep these points in mind, as we will come hack to them later in this chapter.

TABLE 8.4. Evaluations of Self (top half), and Self and Others (bottom half) as a Function of Self-Esteem

_____________________________________________________________________________________________

Note: Values could range from 1 (not at all true of me or others) to 7 (very true of me or others). HSE = high self-esteem (top 1/3 of self-esteem distribution); LSE = low self-esteem (bottom 1/3 of self-esteem distribution). Within each pair, means with different subscripts differ at p < .05 or less.

High self-esteem people not only appraise themselves in more positive terms than do low self-esteem people, they also appear to be more sure of who they are. Campbell and her colleagues (Campbell, 1990; Campbell & Lavallee, 1993) have claimed that high self-esteem people are more apt to possess clearly defined and temporally stable self-views than are low self-esteem people (see also, Baumgardner, 1990; Setterlund & Niedenthal, 1993; Pelham, 1⁹91a). This assertion is based on evidence that low self-esteem people (1) show greater changes in their self-evaluations from one day to the next, (2) take longer to make decisions regarding their attributes, (3) report being less certain of where they stand on various attributes, and (4) display greater inconsistency in their self-evaluations than high self-esteem people. Because people’s ideas of themselves are often used as behavioral guides, the self-concept confusion low self-esteem people show can have important consequences (Baumgardner, 1990; Campbell, 1990; Setterlund & Niedenthal, 1993). For example, low self-esteem people may be more willing than high self-esteem people to accept self-discrepant feedback.

Sociological models provide another perspective on the nature and origins of self-esteem. Building on Cooley’s (1902) model of the “looking-glass self” (discussed in Chapter 3), and Mead’s (1934) ideas about perspective taking and the generalized other (discussed in Chapter 4), sociological models assume that self-esteem is influenced by societal factors. If we think we are highly regarded and valued by society at large, then we have high self-esteem. From this perspective, sociological variables, such as occupational prestige, income, education, and social status (e.g., race, religion, and gender) are assumed to affect self-esteem.

In fact, the evidence supporting such associations is weak. The successful, the affluent, the well educated, and the socially privileged do not have higher self-esteem than people who are less advantaged in these areas (Crocker & Major, 1989; Wylie, 1979). Indeed, members of stigmatized and minority groups sometimes report higher levels of self-esteem than do those who are more privileged (Rosenberg, 1979).

Group pride may explain why self-esteem is not lower in socially disadvantaged groups. As discussed in Chapter 2, minority groups are currently encouraged to view their minority status as a badge of honor rather than a stigma. This perspective is embodied in the Black pride movement, the Gay pride movement, and other similar social movements.

Group pride, in turn, may affect self-esteem. According to social identity theory (Tajfel & Turner, 1986), self-esteem depends, in part, on our group memberships or social identities. People who evaluate their social groups positively enjoy greater self-esteem than those who evaluate their social groups negatively. Crocker and her colleagues (Crocker, Luhtanen, Blame, & Broadnax, 1995; Luhtanen & Crocker, 1992) tested these ideas and found that self-esteem (as measured by the Rosenberg scale) is positively correlated with a measure of collective self-esteem (the degree to which people evaluate their various social groups in favorable terms). Although this correlation doesn’t prove that positive group evaluations promote high self-esteem (it is just as likely that high self-esteem leads people to evaluate their groups positively), it does establish that personal self-esteem and collective self-esteem are related.

Crocker and Major (1989) have offered another explanation for why members of socially disadvantaged groups do not have low self-esteem (see also, Rosenberg, 1979). These theorists have suggested that socially disadvantaged groups protect themselves from prejudice and discrimination (1) by attributing negative feedback to prejudice against their group rather than to themselves; (2) by selectively comparing their own outcomes with other ingroup members rather than with members of the majority group; and (3) by devaluing attributes on which their group is deficient and exaggerating the importance of attributes on which their group excels.

Crocker, Voelkl, Testa, and Major (1991) demonstrated how the first of these processes can insulate socially disadvantaged groups from the negative effects of prejudice and discrimination. In their study, African-American college students completed a questionnaire regarding their attitudes, values, and personal qualities. These students were then told that their responses would be shown to another person (who in all cases was said to be Caucasian), and they would learn whether the other person liked them or not. Finally, students were led to believe that the other person could see them through a one-way mirror (in which case the other person would be aware of their race) or could not see them through a one-way mirror (in which case the other person would be unaware of their race).

In the unseen condition, African-American students felt good about themselves when they were told the other person liked them, and they felt bad about themselves when they were told the other person did not like them. These effects did not occur in the seen condition, however. Here, the students’ feelings toward themselves did not change as a function of the feedback they received. Additional analyses suggested that this occurred because the students attributed the evaluation they received to their race, not to their personality, when they thought the other person had seen them (for related research, see Ruggiero & Taylor, 1997).

Note: Scale values for negative items have been reversed, so that higher numbers equal more favorable self-evaluations.

Gender is another sociological variable that is thought to influence self-esteem. Here again, however, the effects are weak. Contrary to reports circulating in the popular press, research does not show that females have lower global self-esteem than men (Feingold, 1994; Maccoby & Jacklin, 1974; Pliner, Chaiken, & Flett, 1990; Wylie, 1979). Sex differences do emerge, however, with respect to how males and females evaluate certain of their attributes and abilities (Beyer, l990; Marsh, 1990). For the most part, these differences mirror cultural stereotypes.

The study I conducted with University of Washington undergraduates (described earlier) is relevant to this issue. There were 73 men and 60 women in this sample. The two groups had similar scores on the Rosenberg (1965) self-esteem scale, hut they did differ in how they evaluated their various characteristics. Table 8.5 reveals the nature of these effects. The table shows that the 26 items fell into four factors, which I have labeled competence, popularity/attractiveness, kindness, and athleticism. There were no sex differences for the items that tapped perceived competence and popularity/attractiveness, but women thought they were kinder than did men, and men thought they were more athletic than did women. Even here, however, the differences were rather small, and both groups evaluated themselves very positively.

The tendency for women to laud their interpersonal qualities fits with evidence that women place particular importance on these attributes. Earlier (in Chapter 2) we noted that women are more apt than men to develop a collectivist or interdependent self-concept, one that emphasizes their connection and relations with others. Men, on the other hand, are more apt to develop an individualistic self-concept, one that emphasizes their achievements and separateness from others (Kashima et al., 1995; Markus & Oyserman, 1989). Josephs, Markus, and Tafarodi (1992) suggested that this tendency may be relevant to understanding sex differences in self-esteem. They hypothesized that men and women base their self-esteem on different attributes: Women base their self-esteem on their interpersonal qualities and men base their self-esteem on their perceived competence and personal achievements. The evidence for this claim is currently mixed (Nolen-Hoeksema & Girgus, 1994), but the notion that self-esteem depends on the way people evaluate themselves in culturally important domains shows how the cognitive and sociological approaches to understanding self-esteem formation can be integrated.

Terror management theory (Solomon et al., 1991) represents another integrated model of self-esteem development. This theory is based on the work of Ernest Becker. Becker (1973) argued (1) that a defining feature of human existence is the capacity to contemplate one’s own death, (2) that an awareness of one’s own mortality creates anxiety and existential terror, and (3) that cultures function to mollify this terror by prescribing a way of life that is meaningful and valuable and offers some hope for immortality. For example, one culture may emphasize achievement, another charity, and another piety as the appropriate way to live. Extending these ideas, Solomon et al. suggested that high self-esteem develops among people who believe they are upholding the standards and ideals emphasized by their particular culture.

Self-esteem is made possible by the development of cultural worldviews, which provide a stable and meaningful conception of the universe, social roles with specific prescriptions for behaviors that are deemed valuable, and the promise of safety and immortality to those who satisfy those prescriptions. Self-esteem is therefore a cultural contrivance consisting of two components: a meaningful conception of the universe combined with the perception that one is meeting the standards for value within that culturally contrived reality. (Solomon et al., 1991, pp. 24—25)

Because it emphasizes the cultural bases of self-esteem, terror management theory represents a sociological model of self-esteem formation.

Psychological research has not only looked at the nature and origins of self-esteem, it has also explored the consequences of self-esteem. The questions here are “When does self-esteem matter?” and “What difference does it make whether you have high self-esteem or low self-esteem?”

One area of research that has received a lot of attention is the role self-esteem plays when people confront evaluative feedback. Some research looks at how global self-esteem (a personality variable) influences how people cope with evaluative feedback (e.g., Baumeister & Tice, 1985; Brown, 1993); some research explores how evaluative feedback influences feelings of self-worth (e.g., Leary et al., 1995; MacFarland & Ross, 1982); and other research examines how a presumed need to feel good about ourselves (i.e., a self-enhancement motive) guides the way people deal with evaluative feedback (e.g., Steele, 1988; Tesser, 1988).

Our focus will be on the first of these research areas. We will concentrate on understanding how individual differences in self-esteem influence the way people respond to evaluative feedback. The first thing to note is that self-esteem generally has very little effect on how people deal with positive feedback (Brown & Dutton, 1995b; Campbell, 1990; Zuckerman, 1979). With few exceptions (to be discussed later), everyone likes to succeed and feels good when they do.

Where self-esteem matters most is when people confront negative outcomes, such as failure in the achievement domain, interpersonal slights or rejection, and even criticism or disagreement from friends. As a first step toward understanding these effects, consider the following vignettes.

Your boss asks you to prepare a report offering a recommendation on some project. After giving the project considerable thought, you decide it should be approved. You carefully prepare a report outlining your position and turn it over to your boss. Your boss reads your report and rejects your recommendation.

It is nearing lunch time and you decide to go out and get something to eat. You notice three of your co-workers talking. At 12:00, they leave together to have lunch without inviting you.

How would these experiences affect you? Would they make you feel sad and unhappy? Angry and upset? Would they affect the way you feel about yourself? Your answers may reveal a lot about your level of self-esteem. Experiences like these hurt low self-esteem people a lot; they make them feel humiliated and ashamed of themselves and they lead them to think they are worthless and unloved. This is not true of high self-esteem people. High self-esteem people feel sad and disappointed when they encounter negative feedback, but they do not feel humiliated and ashamed of themselves. They do not take failure as personally as do low self-esteem people.¹

An investigation by Brown and Dutton (1995b) documents some of these effects. In the first part of this study, we had a large sample of college students complete the Rosenberg (1965) self-esteem scale. Those who scored in the upper third were designated as having high self-esteem, and those who scored in the lower third were designated as having low self-esteem. (We did not test those who fell between these extremes, because it is difficult to tell whether their self-esteem is high or low.) These participants were then given a test that allegedly measured an important

¹I am using the term failure here in a generic sense to refer to any situation that involves negative self-relevant feedback. This would include not only achievement-related failure but also defeat in the athletic domain and a wide range of interpersonal outcomes, including rejection, criticism from others, and feeling underappreciated or ignored.

intellectual ability. By varying the difficulty of the problems they received, it was possible to lead half of the participants to succeed on the task and to lead the other half to fail.After learning how they had done, the participants completed an eight-item emotion scale. Four of the items (happy, glad, unhappy, sad) represented very general emotional responses to a positive or negative outcome. The other four items (proud, pleased with myself, ashamed, and humiliated) referred specifically to how people feel about themselves. These latter emotions are examples of what we have called feelings of self-worth.

The left-hand side of Figure 8.1 displays the results for the four general emotions. These data show that participants felt sadder after they had failed than after they had succeeded, and that this was just as true of high self-esteem participants as of low self-esteem participants. The situation is different when we look at how participants felt about themselves after learning they had succeeded or failed (see the right side of Figure 8.1). Here we do find an effect of self-esteem. Low self-esteem participants felt good about themselves when they succeeded but they felt bad about themselves when they failed. This was much less true of the high self-esteem participants; how they felt about themselves did not depend so much on whether they had just succeeded or failed.

These data make several points. It’s okay to feel unhappy and disappointed when you fail at something. That is an understandable response to failure, and both high self-esteem people and low self-esteem people show it. But only low self-esteem people feel bad about themselves when they fail. They take failure very personally; it humiliates them and makes them feel ashamed of themselves. High self-esteem people do not show this effect; they do not feel bad about themselves when they fail.

There’s another way to look at these data. Low self-esteem people’s feeling toward themselves are very conditional. If they succeed they feel good about themselves; if they fail they feel bad about themselves. This is a very precarious approach to emotional life. For low self-esteem people, “you’re only as good as your latest outcome.” High self-esteem people do not seem to live this way. How they feel about themselves doesn’t depend so much on what they have just accomplished (Baldwin & Sinclair, 1996).

FIGURE 8.1. Emotional responses to success and failure as a function of emotion type and self-esteem. The data show that both self-esteem groups felt sad when they failed (left-hand panel), hut only low self-esteem participants felt bad about themselves when they failed (right-hand panel).

Kernis (1993; Kernis, Cornell, Sun, Berry, & Harlow, 1993) has noted that this may not be true of all people who score high on self-esteem scales. By measuring self-esteem on repeated occasions, he has identified people who have unstable high self-esteem; these individuals report having high self-esteem, hut their feelings of self-worth fluctuate from day to day.

Unstable high self-esteem represents a kind of pseudo or defensive high self-esteem. Rather than being truly secure in their self-love, people with unstable high self-esteem feel good about themselves only when things are going well. In this sense, unstable high self-esteem appears to be a disguised form of low self-esteem—one in which the person’s feelings of self-worth are highly conditional on recent achievements and events. The comedian David Letterman aptly describes the experience:

Every night you’re trying to prove your self-worth. It’s like meeting your girlfriend’s family for the first time. You want to be the absolute best, wittiest, smartest, most charming, best-smelling version of yourself you can possibly be. That’s how I feel every night I go down there to the Ed Sullivan theater. If I can make these 500 people enjoy the experience, and have a higher regard for me when I’m finished, it makes me feel like an entire person. If I’ve come short of that, I’m not happy. How things go for me every night is how I feel about myself for the next 24 hours. (David Letterman, Parade Magazine. May 26, 1996, p. 6).

A number of factors could explain why low self-esteem people feel worse about themselves when they fail than do high self-esteem people. One possibility is that the two self-esteem groups evaluate their performances differently. Given comparable levels of performance, low self-esteem people might be more inclined than high self-esteem people to view it as a failure. Another possibility is that low self-esteem people might be more apt to attribute a poor performance to low ability.

These effects do occur (Blaine & Crocker, 1993; Campbell & Fairey, 1985; Jussim, Coleman, & Nassau, 1987; Shrauger, 1972; Zuckerman, 1979), but they do not fully explain why low self-esteem people feel so bad about themselves when they fail (Brown & Dutton, 1995b; Dutton & Brown, 1997). Instead, it seems that failure means something very different to a low self-esteem person than to a high self-esteem person. To a low self-esteem person, failure means you are globally inadequate—really, that you are a bad person. To a high self-esteem person, failure simply means that you didn’t do well at some task or that you lack some ability.

An investigation by Dutton (1995) illustrates these effects. After leading participants to succeed or fail on an intellectual task, Dutton had participants evaluate themselves in four different areas: (1) the specific ability measured by the test (“How high/low are you in this ability?”); (2) general intelligence (“How intelligent/unintelligent are you?”); (3) interpersonal qualities (“How kind/phony are you?”); and (4) general perceptions of one’s worth as a person (“Overall, how good/bad a person are you?”).

Figure 8.2 presents the results of this investigation. The first panel shows that both self-esteem groups thought they were lower in the specific ability when they failed than when they succeeded. This is perfectly reasonable. If you have just done poorly on a test, it is reasonable to assume that you lack ability in this area. Now look at the second panel, which shows participants’ perceptions of their general intelligence. Notice the effects of self-esteem here. After failing a test, low self-esteem participants doubt their overall intelligence. They don’t simply assume that failure means they lack a specific ability; they assume it means they lack intelligence altogether. High self-esteem participants do not draw this inference. Although they accept that failure means they lack a particular ability, it doesn’t lead them to believe they are generally lacking in intelligence.

The results are even more dramatic when we look at how failure influenced participants’ perceptions of their interpersonal qualities (e.g., “How kind and warm are you?”). After failing a test of their intellectual ability, low self-esteem participants tended to belittle their social qualities. It’s as if they were saying, “Not only am I bad at this test and generally unintelligent, but come to think of it, I’m not a very nice person either.” High self-esteem participants did not show this tendency. Instead, they tended to compensate for failure by slightly exaggerating their perceived social skills (see also, Baumeister, 1982a; Brown & Dutton, 1995b; Brown & Smart, 1991).

F1GURE 8.2 Self-evaluations as a function of self-esteem, success/failure, and the generality of the evaluation. The data show that low self-esteem people overgeneralize from failure. Failure makes them believe not only that they lack the specific ability measured by the test but also that they are less intelligent, not as socially skilled, and not as good a person. This type of overgeneralization does not occur among high self-esteem people.

(Adapted from Dutton, 1995. unpublished raw data, University of Washington, Seattle.)

Finally, we can look at how test performance affected participants’ perceptions of their general worth as a person (see Panel 4 in Figure 8.2). In comparison with success, failure led low self-esteem participants to disparage their overall worth, not good people, but it did not have this effect on high self-esteem participants (for similar results, see Brown & Dutton, 1995b; Epstein, 1992; Heyman, Dweck, & Cain, 1992; Kernis, Brockner, & Frankel, 1989; Sanbonmatsu, Akimoto, & Moulin, 1994; for related work with depressives, see Beck, 1967; Carver & Ganellen, 1983; Carver, Ganellen, & Behar-Mitrani, 1985; Wenzlaff & Grozier, 1988).

These patterns represent very different responses to failure. Failure hits low-self-esteem people very hard. It makes them feel bad about themselves and it humiliates them. Failure does not have this effect on high self-esteem people. High self-esteem people feel disappointed when they fail, and they may accept that failure means they lack a specific ability. However, they do not treat failure as a global indictment of their character, and they do not feel humiliated and ashamed of themselves.

In my judgment, this is the key difference between having high self- esteem and low self-esteem. The problem with having low self-esteem is that your feelings of self-worth are very conditional. If you succeed, you think you are a good person and feel proud of yourself; but if you fail, you think you are a bad person and feel ashamed of yourself. This is not so for high self-esteem people. They can fail and still feel good about themselves. This capacity to fail without feeling bad about yourself is arguably the chief benefit of having high self-esteem.

Suppose you were devastated by failure, rejection, disappointment, and the like. How would it affect your behavior? Would you readily seek out opportunities to expose yourself to these negative outcomes, or would you become rather self-protective, choosing to avoid circumstances that put you at risk? Accumulating research shows that low self-esteem people generally adopt the second strategy. They become self-protective, choosing to avoid situations that might lead to negative self-relevant feedback in favor of safe, though often less rewarding outcomes (Baumeister et al., 1989; Tice, 1993).

Let’s look first at the relation between self-esteem and risk-taking behavior. The first thing to note is that people generally prefer a sure gain to a speculative, hut potentially more profitable, outcome (i.e., “A bird in the hand is worth two in the bush”). For example, when given a choice between winning $800 or having an 85 percent chance of winning $1,000, people prefer the $800 (despite the fact that the expected utility of the second option is greater) (Tversky & Kahneman, 1981).

Josephs, Larrick, Steele, and Nisbett (1992) showed that low self-esteem people are especially apt to avoid taking risks. These researchers explained their findings in terms of self-protection (see also Larrick, 1993). They noted that pursuing a risky choice of action can not only have negative financial consequences, it can have negative psychological consequences, insofar as a risk that does not pan out casts doubt on one’s decision-making ability and judgment. Because low self-esteem people are especially bothered by negative self-relevant feedback, they pursue a psychologically safer (but potentially less rewarding) strategy.

Note, then, that it is not the fear of financial loss that leads low self-esteem people to be more risk averse. Rather, it is the desire to protect themselves from knowing they have made a bad decision. In support of this interpretation, Josephs et al. (1992) found that low self-esteem people were more risk averse than high self-esteem people only when they anticipated learning whether their choices had been good or bad. If there was no chance they would find out they had made a bad decision, they were not more risk averse.

Research on self-handicapping also shows that low self-esteem people are highly concerned with self-protection. As discussed in Chapters 3 and 7, self-handicapping refers to the fact that people sometimes create impediments to their own success if doing so allows them to preserve an image of competency. For example, a student who fails to study for an important exam may he self-handicapping. Self-handicapping simultaneously serves two functions: (1) it offers self-protection from the pain of failure (rather than attributing failure to low ability, the student can blame a lack of preparation), and (2) it offers self-enhancement given success (a student who succeeds without studying can claim to have superior ability).

Building on earlier theoretical work by Arkin (1981), Tice (1991) hypothesized that low self-esteem people are drawn to self-handicapping as a means of self-protection. When they engage in self-handicapping, it is primarily to protect themselves from the pain of failure by avoiding the implication that they have low ability. In contrast, high self-esteem people are drawn to self-handicapping as a means of self-enhancement. When they engage in selfhandicapping, it is to augment the perception that they have high ability by succeeding despite a self-imposed impediment.

Tice (1991) tested these ideas in a series of studies. In one investigation, participants were told they were going to take a test that measured an important intellectual ability. In the self-protection condition, participants were told that the test could clearly detect low ability but could not clearly identify high ability. These instructions were reversed in the self-enhancement condition. Here, participants were told that the test could clearly identify high ability but could not clearly detect low ability.

FIGURE 8.3. Seconds spent practicing before taking a test that could either clearly disclose high ability (but not low ability) or clearly disclose low ability (but not high ability). The data show that low self-esteem participants handicapped themselves by not practicing when a poor performance would reveal low ability and that high self-esteem participants handicapped themselves by not practicing when a good performance would reveal high ability. These findings support the claim that low self-esteem people use self-handicapping as a form of self-protection and that high self-esteem people use self-handicapping as a form of self-enhancement.

Tice then gave participants time to practice before taking the test (the assumption being that a lack of practice constitutes a form of self-handicapping). Figure 8.3 presents the results of this investigation. The figure shows that low self-esteem participants handicapped themselves by not practicing when they believed that failure could clearly disclose low ability, whereas high self-esteem participants handicapped themselves by not practicing when they believed that success could clearly reveal high ability. These findings suggest that low self-esteem people use self-handicapping to avoid the perception that they have low ability (i.e., a form of self-protection), and that high self-esteem people use self-handicapping to augment the perception that they have high ability (i.e., a form of self-enhancement). Rhodewalt, Morf, Hazlett, and Fairfield (1991) provided evidence that these self-handicapping strategies do, in fact, cushion the blow of failure among low self-esteem people and increase the glow of success among high self-esteem people.

Research has also examined how self-esteem affects task performance and task persistence. In the absence of any prior outcome (i.e., under control conditions) or following prior success, self-esteem has virtually no effect on task performance. But after prior failure, low self-esteem people perform worse than high self-esteem people (Brockner, 1979; Brockner et al., 1983; Shrauger & Sorman, 1977; Shrauger & Rosenberg, 1970). Indeed, merely contemplating failure appears to undermine the performance of low self-esteem people (Campbell & Fairey, 1985).

There are at least two reasons why this occurs. First, failure leads low self-esteem people to become preoccupied with themselves. They become self-focused, and this, in turn, impairs their performance because their attention is no longer on the task at hand (Brockner, 1979; Brockner & Guare, 1983; see also, Dweck & Leggett, 1988 for related work). Second, they withdraw effort. Following prior failure, low self-esteem people persist less at a task than do high self-esteem people (McFarlin, 1985; McFarlin, Baumeister, & Blascovich, 1984; Sandelands, Brockner, & Glynn, 1988; Shrauger & Sorman, 1977). Persistence after failure isn’t always adaptive, of course (Baumeister & Tice, 1985; McFarlin et al., 1984), but low self-esteem people seem particularly insensitive to the conditions under which it will and will not pay dividends (Janoff-Bulman & Brickman, 1982; McFarlin, 1985; Sandelands et al., 1988).

Low self-esteem people continue to be self-protective even after they have performed some task or activity. In Chapter 3 we discussed social comparison processes. Comparing ourselves with others can be quite risky when a highly evaluative dimension is involved. When we compare ourselves with someone else, we run the risk of finding out that we are worse off than they are. Because low self-esteem people are highly concerned with self-protection, they ought to be reluctant to take this gamble unless they are quite certain they will learn good things about themselves.

Wood et al. (1994) tested this hypothesis and found considerable support for it. These investigators first gave participants positive or negative feedback on an alleged test of occupational success. All participants were then given the opportunity to engage in social comparison. Low self-esteem participants eagerly sought social comparison when they thought it would be favorable (after receiving positive feedback), but they actively avoided social comparison when they thought it would be unfavorable (after receiving negative feedback). This cautious, “play it safe” strategy provides further evidence that self-protection is a dominant motivational force in the lives of low self-esteem people (for additional research relating self-esteem and social comparison processes, see Aspinwall & Taylor, 1993; Gibbons & Gerrard, 1989, 1991).

Negative

beliefs

about

ability

Expect

to fail

Generalize

to other

areas of

self-

concept

Fail to compensate through

self-

affirmation

Emotional

distress

Figure 8.4 Schematic model of a cognitive approach to understanding why low self-esteem people suffer greater emotional distress than high self-esteem people when they fail.

To this point, we have seen that failure humiliates low self-esteem people and leads them to believe that they are globally inadequate. We have also seen that low self-esteem people are especially self-protective and risk averse, even to the point that they may create an impediment to success if doing so means they can avoid confronting the fact that they lack ability. Finally, we have seen that low self-esteem people perform worse and persist less after failure and that they choose to compare themselves with others only when they are reasonably certain that they will learn good things about themselves.

Two broad classes of theories have been developed to explain these findings. One focuses on cognitive factors; the other focuses on affective considerations. We will conclude this section by considering each of these in turn. First, we will look at how these theories explain the fact that low self-esteem people are more adversely affected by failure; then we will consider how they explain the greater risk aversion that low self-esteem people show.

Figure 8.4 presents a cognitive model of emotional distress. The model begins by assuming that low self-esteem people doubt their ability. This, in turn, leads them to expect to fail and to attribute failure to low ability when they do. Failure then spreads to other areas of the self-concept.

Low self-esteem people then fail to compensate for this overgeneralization. Steele’s (1988) self-affirmation theory offers one explanation for why this occurs. As discussed in Chapter 5, self-affirmation theory argues that people can neutralize the effects of failure in one domain by emphasizing their virtues in other, unrelated, domains (see also, Tesser & Cornell, 1991). A student, for example, may offset failure in the classroom by emphasizing her interpersonal virtues. Earlier (see Figure 8.2) we noted that high self-esteem people tend to show this response to failure, but low self-esteem people do not (Baumeister, 1982a; Brown & Dutton, 1995b; Brown & Smart, 1991).

Steele and his colleagues have reasoned that low self-esteem people fail to compensate for failure because they have few positive beliefs about themselves to call upon as resources (Josephs et al., 1992; Spencer, Josephs, & Steele, 1993; Steele et al., 1993). From this perspective, the key difference between high self-esteem people and low self-esteem people lies in how they think about their various abilities and attributes. Low self-esteem people believe they have fewer positive qualities than do high self-esteem people, and this is why they experience greater distress when they fail.²

In tracing self-esteem differences to the way people think about themselves, cognitive models present a very rational, logical explanation for why low self-esteem people feel worse when they fail than do high self-esteem people. Affective models offer a somewhat different perspective (Brown, 1993; Dutton & Brown, 1997). These models assume that, early in life, low self-esteem people come to feel bad about themselves whenever they make a mistake or otherwise do something wrong. Over time, these feelings come to be evoked whenever failure is encountered. But the feelings are not cognitively mediated; they are largely automatic, visceral, and irrational.

By way of illustration, imagine that a young child spills his milk and, for whatever reason, becomes extremely distraught. The process is not a cognitive one. The child doesn’t think “I spilled my milk. This means I lack coordination. Being coordinated is a socially desirable quality. I have few other qualities to offset the impact of this deficiency. Therefore, I’m bad.” The reaction is more undifferentiated and automatic than this; the child simply assumes that he is had when he spills his milk. The affective model argues that low self-esteem people still respond to failure in just this way. Failure automatically makes them feel had about themselves.

² Steele's (1988) self-affirmation theory is related to Linville’s (1975) theory of self-complexity (see Chapter5). The key difference is that self-affirmation theory maintains that the pain of failure is reduced if the person possesses many other positive identities, whereas self-complexity theory claims that the emotional effects of success and failure are less extreme among people who possess many identities, regardless of whether these identities are positive or negative.

This account draws a parallel between early childhood experiences and later responses to failure. Epstein (1980) also noted these parallels.

People with high self-esteem, in effect, carry with them a loving parent who is proud of their successes and tolerant of their failures. Such people tend to have an optimistic view about life, and to be able to tolerate external stress without becoming excessively anxious. Although capable of being disappointed and depressed by specific experiences, people with high self-esteem recover quickly, as do children who are secure in their mother’s love. In contrast, people with low self-esteem carry within them a disapproving parent who is harshly critical of their failures, and register only short-lived pleasures when they succeed. Such people are apt to be unduly sensitive to failure and to rejection, to have low tolerance for frustration, to take a longtime to recover following disappointments, and to have a pessimistic view of life. The picture is not unlike that of children who are insecure in their parent's love. (Epstein, 1980, p. 106)

Cognitive and affective models also offer different explanations for why low self-esteem people are more self-protective and risk averse than are high self-esteem people. Expectancy-value models of behavior provide a useful framework for understanding these differences. As discussed in Chapter 6, these models assume that freely chosen behavior depends on two factors: a person’s expectation that she can achieve some outcome, in conjunction with the value the person places on obtaining versus not obtaining the outcome.

Cognitive theories trace self-esteem differences in behavior to the expectancy component of the expectancy-value model. They assume that low self-esteem people are reluctant to take risks because they have little confidence in their ability to succeed. Affective models emphasize the value component in the expectancy-value model. They assume that the negative incentive value of failure (the pain of failure) is greater for low self-esteem people than for high self-esteem people and that this explains why low self-esteem people are more self-protective.

One way to think about these divergent perspectives is in terms of confidence and consequence. The cognitive model assumes that a lack of confidence guides the behavior of low self-esteem people. Low self-esteem people play it safe because they lack confidence that they will succeed. Presumably, if they thought they were better at things (e.g., thought they had higher ability), they would not be risk averse. The affective model assumes that consequence, not confidence, is the key variable to consider. Low self-esteem people are risk averse not because they don’t think they can succeed but because they are afraid to fail. Literally, they are afraid to try.

Brockner (1984) has suggested that the combination of these factors makes low self-esteem people especially vulnerable to social influence.Lacking confidence in themselves, and being overly sensitive to rejection and criticism, low self-esteem people conform to the wishes of others. This may explain why low self-esteem is linked to many negative behaviors in adolescence, such as substance abuse and unsafe sexual practices (Hawkins, Catalano, & Miller, 1992). These behaviors often arise as a result of peer pressure, which low self-esteem people may find particularly difficult to withstand.

The data we have reviewed in this chapter are relevant to two long-standing controversies The first is known as the “self-enhancement versus self-consistency debate.” Self-consistency theories (see Chapter 3) assume that people strive to maintain consistency between their beliefs, attitudes, and behavior, and that inconsistency creates an uncomfortable state of tension that people are driven to reduce. Accordingly, consistency theories predict that people with positive self-views will be less accepting of and more disturbed by negative feedback than will people with negative self-views (because negative feedback is more inconsistent with their positive self-image).

Self-enhancement theories make a different prediction. These theories assume that people strive to feel good about themselves. It is also assumed that this need increases the longer it goes unfulfilled (much as a need for food increases the longer one goes without eating). Accordingly, self-enhancement theories predict that people with negative self-views will be less accepting of and more disturbed by negative feedback than will people with positive self-views.

Shrauger (1975) surveyed the literature and found support for each theory, depending on whether cognitive or emotional reactions to negative feedback were assessed (see also, Swann, Griffin, Predmore, & Gaines, 1987). People’s cognitive reactions to negative feedback conform to the self-consistency model: People with negative self-views are more accepting of negative feedback (i.e., they are more inclined to view it as accurate and attribute it to themselves) than are people who have positive self-views. In contrast, people’s emotional reactions to negative feedback conform to the self-enhancement model: People with negative self-views are more disturbed by negative feedback than are people with positive self-views.

Although Shrauger’s review concerned the effects of task-specific beliefs and expectancies, his analysis also illuminates how global self-esteem affects responses to negative feedback. In accordance with the self-consistency model, low self-esteem people are more inclined to accept negative feedback than are high self-esteem people (Blame & Crocker, 1993); in accordance with the self-enhancement model, low self-esteem people are more disturbed by negative feedback than are high self-esteem people (Brown & Dutton, 1995b; Dutton & Brown, 1997).

The situation is a bit more complex when we look at responses to positive feedback. Self-consistency theory predicts that low self-esteem people will be uncomfortable with positive feedback because it is inconsistent with how they think and feel about themselves. Although some studies find support for this prediction (e.g., Brown & McGill, 1989; Marecek & Mettee, 1972), most do not. Most studies find that self-esteem has little effect on how people respond to positive feedback (e.g., Brown & Dutton, 1995b; Campbell, 1990; Shrauger & Lund, 1975).

Swann’s (1990, 1996) research on self-verification processes provides an important exception. As discussed in Chapter 3, self-verification theory is concerned with interpersonal behavior. The theory assumes that people “want others to validate and confirm their [self-views], even when those [self-views] are negative” (McNulty & Swann, 1984, p. 1013). This assumption leads Swarm to predict that people with negative self-views will be uneasy receiving positive feedback from others.

Swarm and his colleagues (De La Ronde & Swarm, 1993; Swann, 1996) have used these ideas to understand the behavior of low self-esteem people. They have argued that low self-esteem people are torn between two competing motives. On the one hand, they want to feel good about themselves, and this leads them to seek and embrace positive feedback from others. On the other hand, they do not want others to think too well of them, because they fear not being able to live up to these expectations. These opposing tendencies place low self-esteem people in a crossfire: They sometimes avoid positive feedback from others (Swarm, Pelham, & Krull, 1989) despite evidence that they are especially hurt by interpersonal rejection (Dittes, 1959; Jones, 1973; Smith & Smoll, 1990). In extreme cases, self-verification needs may lead low self-esteem people to seek out or remain in negative interpersonal relationships (Swann, 1996).

Closely related to the “self-enhancement—self-consistency” debate is the question of whether self-esteem is best treated as a global personality variable or in terms of how people evaluate themselves in specific domains. In keeping with a general movement in the field of personality, many contemporary theorists (e.g., Bandura, 1986; Marsh, 1990; Swarm, 1990) have argued that specific self-evaluations are the better predictor of behavior (broadly defined) than is global self-esteem. Some have even gone so far as to claim that global self-esteem is a fiction (Gergen, 1971) or is of limited value:

[My] research has led me increasingly to the conclusion that general self-concept—no matter how it is inferred—is not a particularly useful construct. General self-concept does not reflect adequately the diversity of specific self-facets. If the role of self-concept [research] is to better understand the complexity of self in different contexts, to predict a wide variety of behaviors, to provide outcome measures for diverse interventions, and to relate self-concept to other constructs, then the specific facets of self-concept are more useful than a global indicator. (Marsh, 1990, p. 100)

The evidence on this point is actually mixed. For example, people’s specific beliefs about their academic ability are a better predictor of school performance than is global self-esteem (Marsh, 1990), but global self-esteem is a better predictor of psychological well-being than are domain-specific self-evaluations (Rosenberg, Schooler, Schoenbach, & Rosenberg, 1995). The evidence is also mixed when it comes to understanding people’s responses to success and failure. Some investigations have found that specific self-views are better predictors of these reactions than is global self-esteem (e.g., Feather, 1969; Marsh, 1990; Swarm et al., 1989); other studies have found the opposite to be true or have reported mixed findings (e.g., Brockner & Hulton; 1978; Dutton & Brown, 1997; Moreland & Sweeney, 1984; Shrauger & Sorman, 1977).

Shrauger’s (1975) affect—cognition distinction appears to illuminate this issue as well. Dutton and Brown (1997) found that task-specific self-evaluations (i.e., expectancies of success and beliefs about one’s competence) guide peopie’s cognitive reactions to evaluative feedback but that global self-esteem guides people’s emotional reactions to evaluative feedback. These findings suggest that both constructs are important and influence different aspects of psychological life.

Throughout this chapter, I have contrasted an affective model of self-esteem with one that emphasizes cognitive factors. In doing so, I have tried to describe the main points of each perspective, without commentary or advocacy. In the final section of this chapter, I will depart from this approach and critically examine the two models.

Let me begin by noting that the cognitive view of self-esteem currently dominates contemporary thinking in personality and social psychology. In part, this is because it fits with the more general cognitive zeitgeist. It is also better articulated than the affective model, and it lends itself better to empirical tests and elaborations. These are all highly desirable features of a theory.

Nevertheless, I personally think the theory fails to adequately capture the nature of self-esteem. This is true with respect to the account it offers for the origins of self-esteem and with respect to its explanation for why low self-esteem people feel so bad when they fail. We’ll begin by considering the first of these issues.

Cognitive models of self-esteem formation adopt an information-integration approach to understanding the origins of self-esteem. They assume that self-esteem develops from a largely rational process. People survey their various characteristics and somehow combine this information into an overall judgment. Stanley Coopersmith (1967), a pioneer in the area of self-esteem research, summarized this approach when he wrote:

[Self-esteem] is based on a judgmental process in which the individual examines his performance, capacities, and attributes according to his personal standards and values, and arrives at a decision of his own worthiness. (Cooper-smith, 1967, p. 7, emphasis added)

I do not think this is so. In my judgment, people don’t examine their various qtialities and decide whether or not to like themselves. People’s feelings toward themselves are more irrational than this. As the French philosopher Pascal noted, “The heart has its reasons, which reason does not know.” Although Pascal was referring to love for others, the same principle applies when we consider how people feel about themselves.

This position was shared by William James. Although his claim that “self-esteem = successes/pretensions” implies that self-esteem is based in cognitive processes, James also believed that self-esteem was not so logical. “There is a certain average tone of self-feeling,” James wrote, “which each one of us carries about with him, and which is independent of the objective reasons we may have for satisfaction or discontent” (James, 1890, p. 306). One way to reconcile this apparent contradiction is to assume that James’s formula for self-esteem applies to feelings of self-worth (e.g., the pride we feel in a given outcome or achievement) rather than to global self-esteem itself.

Another problem I see with the cognitive approach is that the self-esteem it describes is conditional and fragile. Self-esteem founded on the belief that we possess certain qualities would be unstable (Kernis et al., 1993) and vulnerable to attack. In the event that something happened to undermine our self-evaluations in that domain, our self-esteem would evaporate. If it is to have any value, self-esteem ought to insulate people from just such experiences. High self-esteem people ought to (and do) feel good about themselves even when they find out they are not capable or competent in some domain. This analysis further suggests that high self-esteem is not based on an assessment of one’s constituent qualities.

The cognitive model also assumes a high degree of cognitive sophistication. As such, the model predicts that self-esteem doesn’t develop until at least middle childhood, when the cognitive abilities necessary to make the various judgments the model describes have developed. Susan Harter, a developmental psychologist, makes this point explicitly, asserting that global self-esteem is “a complex, cognitive construction that does not emerge until approximately the mental age of eight” (Harter, 1986, p. 145). This characterization is at odds with evidence that children show differences reflective of self-esteem at a very young age and that these differences predict self-esteem later in life (Cassidy, 1990; Sroufe, 1983; Sroufe, Carlson, & Shulman, 1993).

Finally, the cognitive model begs the question of what determines self-evaluations in the first place. Consider, for example, people’s ideas about how attractive they are. At all ages, and for both sexes, perceived attractiveness is closely related to self-esteem (Harter, 1993; Pliner et al., 1990). People who like the way they look, like themselves (and people who like themselves, like the way they look). The cognitive approach assumes that the causal arrow goes from perceived attractiveness to self-esteem. People somehow come to regard themselves as attractive or unattractive, and this decision affects their level of self-esteem. What this approach leaves unanswered is the question of why some people regard themselves as attractive to begin with.

One possible solution would be to assume that people correctly perceive how attractive they really are. But as we noted in Chapter 3, this is not the case. People’s ideas about their attractiveness are not strongly tied to what others think; nor is actual attractiveness related to self-esteem (Feingold, 1992). The same is true for virtually all highly evaluative attributes. High self-esteem people think they are more competent, intelligent, talented, and well liked than do low self-esteem people, but this is not actually the case. On average, high self-esteem people are no better in these areas than are low self-esteem people, and people who truly possess these qualities do not have higher self-esteem than people who lack them.

What we find, then, is that although self-esteem is strongly related to what people think they are like, it is virtually unrelated to what people are really like. This pattern poses a problem for the cognitive model. The problem is this: If people’s ideas about themselves aren’t based on what they are really like, where do they come from?

An affective model of self-esteem explains these findings by assuming that self-esteem develops early in life and then functions as a lens through which people view their characteristics and experiences. People who feel good about themselves evaluate themselves positively—they like the way they look, they appreciate their talents, and they believe they are warm, friendly, and liked by others. The relation is a top-down one (from global self-esteem to self-evaluations), rather than a bottom-up one (from self-evaluations to global self-esteem) (Brown, 1993; Brown, Dutton, & Cook, 1997).

Cognitive explanations for why low self-esteem people feel so bad about themselves when they fail may also be incomplete. According to these models, ¹ow self-esteem people (1) think they have low ability; (2) expect to fail; (3) are accepting of failure when it occurs; (4) generalize from the failure experience to other areas of the self-concept; and (5) fail to compensate for failure because they don’t think they do other things well.

At one time, I was one of the many theorists who endorsed this position (e.g., Brown, Collins, & Schmidt, 1988; Brown & Gallagher, 1992; Brown & Smart, 1991). But now I think it is mistaken. One problem is that the negativity it assumes on the part of low self-esteem people doesn’t really exist. Look again at Table 8.4. Notice that low self-esteem people do not describe themselves in negative terms. They don’t claim to be unintelligent, incompetent, and homely losers. They actually regard themselves quite favorably, believing they are smarter, kinder, and quite a bit better liked than are most other people and that they are certainly every bit as attractive, talented, and sexy. These findings are at odds with a cognitive model that takes as its starting point the idea that low self-esteem people evaluate themselves negatively. It doesn’t make sense to claim that low self-esteem people are so adversely affected by failure because they think they can’t do anything well and have many negative qualities, when in fact they regard themselves as smarter than most other people and not at all incompetent.

It might he argued that this lack of negativity occurs only among low self-esteem college students. I am not aware of any evidence that college students score higher in self-esteem than do other groups, but true self-deprecation does occur in some populations (e.g., the severely depressed). Nevertheless, the fact remains that self-esteem differences are found with low self-esteem people who do not evaluate themselves in highly negative terms. It must be because these effects are not due to negative thinking per se.

The larger point here is this. Cognitive models assume that the defining feature of low self-esteem is self-criticism. Many low self-esteem people do not fit this pattern. They think they have many fine qualities, yet they feel humiliated and ashamed of themselves when they fail (Bednar, Wells, & Peterson, 1989). The problem is not a lack of positive self-evaluations; rather, it is an almost inexplicable feeling that one is a bad person, a feeling that automatically surfaces whenever failure is encountered. Indeed, one can easily imagine a low self-esteem person who says, “Yes, I know I am smart and attractive and can do many things well, but I just don’t feel good about myself, especially when I fail or make a mistake.”

To summarize, I believe the cognitive model paints too rational a portrait of self-esteem formation and functioning. My claim is not that people’s thoughts about themselves are unrelated to how they feel about themselves, it is that self-evaluations aren’t the relevant thoughts. The cognitions that underlie self-esteem occur at a preconscious level; they are part of what Epstein (1990) calls the experiential system, rather than the rational system. These automatic thoughts (Bargh & Tota, 1988; Beck, 1967) or irrational beliefs (Ellis, 1962) are undifferentiated and diffuse; they are vague notions about one’s general worth as a person (e.g., “I am good versus I am bad”), not ideas about particular competencies and attributes.

The analysis I have offered has implications for the treatment of self-esteem. Many self-esteem improvement programs attempt to instill high self-esteem by encouraging people to focus on their positive qualities (McGuire & McGuire, 1996; Mruk, 1995). This assumes that positive self-evaluations produce high self-esteem. The fact that many low self-esteem people already view themselves positively casts doubt on this assumption.

In a similar vein, attributional retraining programs encourage people to attribute failure to factors other than low ability (Seligman, 1991). The implication is that a perceived lack of ability makes people feel bad about themselves. There is certainly some truth in this. But encouraging people to make such attributions also encourages them to tie their feelings of self-worth to their outcomes and perceived competencies. In the long run, this strategy can be maladaptive. There are many things we can’t do well in life, and there is no reason to feel bad about ourselves because of this, unless we make our self-esteem dependent on our self-evaluations.

Instead of treating low self-esteem by building people’s confidence in themselves (“you can do it”), I believe we should be helping people to understand that “It’s OK if you can’t do it; it doesn’t mean you’re a bad person. Unfortunately, doing so is easier said than done. I am not a clinical psychologist, and my recommendations in this area are little more than intuitions. Still, my intuitions tell me that feelings of belonging and mastery (which I regard as the defining features of high self-esteem) are best created in the context of a secure interpersonal relationship in which a person feels unconditionally loved and accepted.

This position parallels a school of therapy Carl Rogers developed in the 1940s. Rogers (1951) believed that the therapist’s role is to accept the client in a nonjudgmental way, to assure the client that he or she is valued as a person. As in the present approach, this strategy assumes that improvements in self-esteem require a focus on global feelings toward oneself rather than cognitively based judgments of one’s competencies or qualities. Ultimately, self-esteem is not a decision but a feeling. It is based not on a dispassionate consideration of what one is but on feelings of affection for who one is.

	HSE	LSE			HSE	LSE
Athletic	5.56_a	4.81_b		Inadequate	1.36_a	2.17_b
Attractive	5.13_a	4.38_b		Incompetent	1.51a	2.14_b
Capable	6.15_a	5.21		Inconsiderate	1.85_a	2.41_b
Compassionate	5.59	5.21		Insensitive	2.13	2.38
Creative	4.97_a	4.40_b		Insincere	2.05	2.50
Friendly	5.79	5.45		Phony	1.54	1.86
Generous	4.95	4.97		Thoughtless	1.82	2.19
Good Looking	5.21_a	4.33_b		Unattractive	1.67_a	2.60_b
Kind	6.05_a	5.52_b		Uncoordinated	1.56_a	2.21_b
Loyal	6.23_a	5.60_b		Unintelligent	1.41_a	1.95_b
Sexy	4.79_a	3.60_b		Unpopular	1.87_a	2.74_b
Smart	5.67_a	5.05_b		Unwise	1.95	2.33
Talented	5.44_a	4.50_b
Well Liked	5.74_a	5.02_b
HIGH SELF-ESTEEM Athletic Attractive Capable Compassionate Creative Friendly Generous Good Looking Kind Loyal Sexy Smart Talented Well Liked	Self 5.56_a 5.13_a 6.15_a 5.59_a 4.97_a 5.79_a 4.95_a 5.21_a 6.05_a 6.23_a 4.79_a 5.67_a 5.44_a 5.74_a	Others 4.44_b 4.31_b 5.26_b 4.56_b 4.44_b 4.92_b 4.10_b 4.08_b 4.87_b 4.13_b 3.64_b 4.59_b 4.59_b 4.64_b	LOW SELF-ESTEEM Athletic Attractive Capable Compassionate Creative Friendly Generous Good Looking Kind Loyal Sexy Smart Talented Well Liked		Self 4.81 4.38 5.62_a 5.21_a 4.40 5.45_a 4.97_a 4.33 5.52_a 5.60_a 3.60 5.05_a 4.50 5.02_a	Others 4.64 4.40 4.93_b 4.38_b 4.45 4.69_b 4.17_b 4.17 4.76_b 4.14_b 3.67 4.62_b 4.52 4.52_b
Inadequate Incompetent Inconsiderate Insensitive Insincere Phony Thoughtless Unattractive Uncoordinated Unintelligent Unpopular Unwise	1.36_a 1.51_a 1.85_a 2.13_a 2.05_a 1.54_a 1.82_a 1.67_a 1.56_a 1.41_a 1.87_a 1.95_a	2.79 2.92 3.41 3.33 3.31 3.72 3.26 3.23 3.00 2.79 3.23 3.21	Inadequate Incompetent Inconsiderate Insensitive Insincere Phony Thoughtless Unattractive Uncoordinated Unintelligent Unpopular Unwise		2.17a 2.14a 2.40a 2.38a 2.50a 1.86a 2.19a 2.60_a 2.21_a 1.95_a 2.74_a 2.33_a	2.88b 2.90_b 3.48_b 3.48_b 3.52_b 3.69_b 3.19_b 3.14_b 3.07_b 2.67_b 3.26_b 2.90

Competence	Males	Females	Popularity/ Attractiveness	Males	Females
Capable	5.84	5.88	Attractive	4.81	4.78
Creative	4.73	4.57	Sexy	4.36	4.02
Smart	5.40	5.33	Well liked	5.36	5.42
Talented	5.08	4.72	Unattractive	5.84	6.01
Inadequate	6.14	6.12	Good looking	4.88	4.65
Incompetent	6.23	6.03	Unpopular	5.75	5.57
Unintelligent	6.30	6.40	Mean	5.17	5.08
Unwise	5.81	5.92
Mean	5.69	5.62

Kindness	Males	Females	Athleticism	Males	Females
Compassionate	5.15	5.68	Athletic	5.60	4.43
Friendly	5.59	5.78	Uncoordinated	6.22	5.77
Generous	4.89	5.32	Mean	5.91	5.10
Kind	5.66	6.05
Loyal	5.77	5.87
Inconsiderate	5.80	6.20
Insensitive	5.51	6.10
Insincere	5.53	6.05
Phony	6.08	6.45
Thoughtless	5.88	6.15
Mean	5.59	5.97