The Attack on Self-Esteem

(Da Capo Books, 2014)

The Attack on Self-Esteem

By Alfie Kohn

A new idea is hatched; it catches on; it begins to spread; it inspires a flurry of books and articles, conferences and seminars. And then it fades away. The cycle is common in many fields, but I’m most familiar with how it plays out in education, where the last couple of decades have witnessed a sudden (albeit fleeting) excitement about “outcome-based” and then “brain-based” schooling, about Total Quality Management, multiple intelligences, the “flipped” classroom, and several other hot developments, each of which cooled and was supplanted by the next big thing.

In the 1980s and ’90s, self-esteem took its turn in the procession, with that phrase becoming a rallying cry for many American educators. Trying to improve children’s perceptions of their own worth was described as a crucial contributor to how they fared academically and socially. The tipping point came in 1990 with the much-publicized release of a state-funded task force report in California. School-based programs to raise students’ self-esteem began to sprout up across the country. State and local councils devoted to the cause were formed, newsletters circulated, and classroom curricula disseminated.

The quality of these efforts varied widely; they included carefully planned intervention programs as well as silly rituals in which kids were surrounded by cheerfully reassuring posters and made to chant “I’m special!” Some proponents regarded self-esteem not merely as important but as a cure-all—a “social vaccine” against crime and violence, substance abuse, and other cultural diseases—even though a scholarly monograph commissioned by the California task force failed, rather embarrassingly, to find much data to justify this enthusiasm.1

At some point in the 1990s educators began drifting off to other projects. But the effort did leave one enduring legacy. It didn’t have to do with any of the publications or programs dedicated to supporting self-esteem, though—it was the critical backlash that had sprung up. No sooner did educators express an interest in trying to help kids feel better about themselves than the attacks began. A number of these articles, not surprisingly, were written by conservative thinkers,2 but, as with so many other aspects of parenting and education, their response was mirrored in mainstream media outlets. Hence “The Trouble with Self-Esteem” in US News & World Report (1990), “Hey, I’m Terrific” in Newsweek (1992), “Down from the Self-Esteem High” in the New York Times (1993), “A Full Head of Esteem” in the Washington Post (1995), “Self-Esteem Self-Defeating?” in the Boston Globe (1996), and many more along the same lines.

By now the councils and newsletters and classroom exercises devoted to boosting self-esteem are few and far between. Yet the furious denunciations have continued: Another article called “The Trouble with Self-Esteem,” this one in the New York Times Magazine; “The Self-Esteem Hoax” in the Christian Science Monitor; and “Self-Esteem: Why We Need Less of It” in Time all appeared in 2002. Then, too, there are the criticisms in books, from The Myth of Self-Esteem (1998) to The Self-Esteem Trap (2009); in Doonesbury; in best sellers about children such as NurtureShock; and in a steady stream of blog posts, newspaper columns, and discipline manuals. Like someone still shrieking about the threat of Communism, the attack continues long after its target has largely faded away.

The idea that kids should be helped to regard themselves more favorably continues to arouse such disproportionate contempt—the phrase “using a howitzer to kill a butterfly” comes to mind—that one imagines a psychoanalyst rubbing his chin thoughtfully and musing about the critics’ unconscious motives. Having combed through major American newspapers, newsweeklies, and general-interest magazines, I’ve been unable to find a single article on the topic that doesn’t take this negative view. Granted, there are plenty of self-help books about building self-esteem, but in mainstream periodicals, any time you see that phrase (at least in the context of kids), you can count on its being preceded by the adjective “inflated” or accompanied by a warning about “narcissism.” Indeed, the arguments and rhetoric are so similar from one article or book to the next that you may find yourself wondering if a single person wrote all of them. And yet, just as with polemics about overparenting or the benefits of failure, many of these attacks on self-esteem are delivered in a tone of self-congratulation, as if it took extraordinary gumption to say pretty much what everyone else is saying.


Critics of self-esteem have a tendency to slide into other familiar complaints: They also condemn excessive praise (as if self-esteem and praise were essentially the same thing),3 our alleged reluctance to criticize children, the diminution of competition (and therefore excellence), the ease with which grades and trophies are acquired, and so on. Sometimes all these grievances are bundled together under the heading of “the self-esteem movement” in order to permit more efficient disparagement of everything that rankles traditionalists.

A fair amount of what’s said in these articles and books consists of sweeping generalizations about lazy, entitled kids. The cumulative result of these attacks is that many parents and teachers are now embarrassed to point out that something is damaging to their children’s self-esteem, even when that’s an accurate description and a legitimate concern. Still, there are several questions about the concept that clearly need to be taken seriously. It’s certainly reasonable, for example, to point out that raising someone’s self-esteem isn’t easy to do, and that, even if successful, such efforts aren’t guaranteed to help children excel in school or become better people. So let’s take a moment to review what we know about the subject.

Does self-esteem really matter?

Those who regard the whole concept as something unworthy of serious concern include some psychologists—Jean Twenge, Martin Seligman, and Roy Baumeister, to name three—who argue that it’s merely a by-product of other qualities or that high self-esteem is not particularly beneficial and may even be pathological. However, they tend to make these arguments while rattling off various other complaints about children and parenting that are commonly heard from conservative social critics, suggesting that their views about self-esteem may reflect a larger worldview.4

It’s fair to say there are difficulties with how the concept of self-esteem has been conceptualized and investigated, and cause and effect sometimes remain confused, as is true of many other topics in human behavior. Still, it’s clear to the great majority of psychologists—theorists, researchers, and clinicians—that self-esteem does have predictive value. How people view themselves really does mean something.

Adolescents with low self-esteem, for example, have “poorer mental and physical health, worse economic prospects, and higher levels of criminal behavior during adulthood,” even when other variables such as social class, gender, and depression are held constant. Preadolescents with low self-esteem are more likely to be aggressive a couple of years later and also to engage in “problem eating, suicidal ideation, and multiple health compromising behaviors.” And low self-esteem apparently causes depression more than the other way around.5

People with high self-esteem, meanwhile, are apt to be more satisfied with life, less depressed, and more optimistic.6 (There are other positive associations as well, but it’s not always clear that they’re caused by higher self-esteem.) What’s more, researchers have documented a set of interesting connections to success and failure: First, people who hold a positive view of themselves are more likely to persist at a task even when it’s difficult. At the same time, such people are more likely to recognize when persistence would be futile; they don’t keep trying desperately in a way that proves self-defeating and irrational. Finally, higher self-esteem appears to create resilience so that the experience of failure isn’t as discouraging.7

Does self-esteem promote higher achievement?

Here’s a good example of where the concept was oversold. When you look at the research over the last few decades, the correlation between self-esteem and school achievement isn’t all that impressive. To make matters worse, even when they are related, it’s not clear that the former is responsible for the latter. Part of the problem, however, turns on how broadly or narrowly self-esteem has been defined. A lot of these studies focus “on the capacity of global measures of self-esteem to predict specific outcomes.”8 If you score kids on how they respond to prompts such as “I feel that I have a number of good qualities,” the results aren’t all that useful for predicting how well they’ll do in math. But when you look at how children view their capability in a specific field, that does predict their performance. The effect isn’t huge, but it’s consistent, and it does seem to be causal according to two independent reviews of the research. “There is clear evidence . . . that prior levels of academic self-concept lead to higher levels of subsequent academic achievement beyond what can be explained by prior levels of academic achievement.”9

Perhaps you’ve heard people say that “self-esteem isn’t the cause of achievement; it’s the result.” This has become the mantra of traditionalists. Unfortunately, it represents an enormous oversimplification. I’ve already suggested that the first part of that statement is false, or at least greatly overstated: Some versions of self-esteem do contribute to achievement. But the evidence to support the second part, the proposition that doing well in school raises self-esteem, is “disappointingly weak.”10 Then, too, there’s the question of whether “doing well in school” refers to accomplishments that are meaningful to students. (No “I Believe in Me!” self-esteem unit could possibly be more foolish than the expectation that students will feel good about themselves because they successfully filled out a worksheet or memorized a bunch of facts for a quiz. Ironically, many critics of self-esteem seem to prefer just this sort of schooling.)

People who insist that achievement produces self-esteem rather than the other way around are mostly telling us what they think ought to be the case. After all, even if there’s reason to doubt that A causes B, that doesn’t allow us to conclude that B causes A. It may be that something else altogether (C) causes both A and B, giving the appearance of a direct connection between the two.11 Or it may be that A affects B, which, in turn, affects A. Indeed, there’s good reason to believe that “gains in academic achievement that are facilitated by self-esteem, for example, may further enhance feelings of self-worth, thus setting the stage for additional achievement in school.”12

A group of Australian researchers went a step further. They cautioned that it’s actually counterproductive to ignore how kids feel about themselves (or about what they’re doing) and focus only on how well they’re doing it. “Interventions aimed at enhancing performance may unintentionally undermine self-concept in ways that will eventually undermine the short-term gains in performance,” they pointed out. Their example was a study in which both competitive and cooperative strategies were introduced to improve physical fitness. Both produced temporary benefits, but cooperation improved—and competition undermined—the kids’ beliefs about their physical ability.13 It’s shortsighted to concentrate only on skills and other outcomes. If children feel worse about themselves, which is a typical long-range effect of competition for both winners and losers, then any benefits that do show up aren’t likely to last.14 Thus, self-esteem really does play an important role—even if we’re concerned only about achievement. And if we want to produce people who are also fully functioning, happy, and healthy, it matters even more.

Do programs to boost children’s self-esteem actually work?

When I reviewed the research on this topic in the early 1990s, I didn’t see much evidence of success. But there’s now more reason to be hopeful that school-based interventions can make a difference—provided that (a) the focus is on improving the way children view their aptitude in specific areas, (b) the measure of success matches that focus rather than looking at global self-esteem, and (c) the program isn’t ridiculous.15 A 1998 review of 102 studies pronounced the results “encouraging. Programs seem able to enhance children’s and adolescents’ [self-esteem].” Eight years later, another group of researchers reanalyzed those evaluations—as well as an additional batch of studies—using different statistical techniques. They, too, found “promising” results regarding the “overall positive effectiveness of the interventions.”16

Isn’t it possible to have self-esteem that’s too high?

The short answer: only if you’ve stacked the deck, rhetorically speaking, by defining high self-esteem as something bad. That’s basically what Roy Baumeister did in a 1996 article whose subtitle was “The Dark Side of High Self-Esteem.” His essay—which presented no new data, incidentally—was snapped up by the media and to this day is still triumphantly cited by critics of self-esteem: Aha! Dangerous criminals actually think too well of themselves, not too poorly! Self-esteem is the problem, not the solution. But Baumeister’s conclusion was preordained by his premise. On his article’s very first page, “high self-esteem” and “egotism” were used interchangeably. Since many violent people are egotistical, there must be a risk when self-esteem reaches a certain level. Q.E.D.

Complementing this and other examples of dubious and even offensive reasoning17 was a remarkably simplistic understanding of human psychology. Baumeister basically assumed that we should take people’s sweeping self-congratulatory statements about themselves—“I’m the greatest/smartest/ strongest!”—at face value. Anyone who brags about how amazing he is must have very high self-esteem. This premise is also accepted by like-minded critics such as Jean Twenge (“As any parent of a two-year-old can tell you, most kids like themselves just fine—and make the demands to prove it”18), but it’s unconvincing to anyone who realizes there’s a world of difference between genuinely positive self-regard and arrogant self-satisfaction.

Even people who have never read Freud or other depth psychologists understand that someone who feels compelled to swagger and boast, to flash his credentials or his bling, to tell you how much better he is than everyone else, may well be trying to compensate for the terrifying suspicion that, down deep, he’s really not very impressive at all. Scratch a competitive person and you’ll likely find persistent insecurity and self-doubt. So, too, for narcissists: They “report a grandiose sense of self . . . yet, covertly, they seem to experience symptoms of vulnerability; they are self-doubting.”19 The same is true of those who are aggressive. As the psychiatrist James Gilligan, an expert on criminal violence, has remarked, “I have yet to see a serious act of violence that was not provoked by the experience of feeling shamed and humiliated. . . . The most dangerous men on earth are those who are afraid that they are wimps.”20

Grandiosity, narcissism, and perhaps competitiveness may be understood as strategies for dealing with underlying low self-esteem. But even if this isn’t always true, people with those characteristics are clearly very different from those with high self-esteem. In the words of the late Morris Rosenberg, one of the pioneers in studying this topic, “With self-esteem we are asking whether the individual considers himself adequate—a person of worth—not whether he considers himself superior to others.”21 To that extent, the fact that grandiose people may become aggressive provides no reason to believe there’s a “dark side” of high self-esteem.

Research confirms this. Whether or not high self-esteem and narcissism are positively correlated—and there’s mixed evidence about that22 — it’s clear that “genuine self-esteem and narcissistic self-aggrandizement are distinct constructs.”23 Further, a series of studies with eleven-year-olds, fourteen-year-olds, and undergraduates found that aggression and delinquency were negatively related to self-esteem, meaning that we have more reason to worry when it’s low than when it’s high.24 The same pattern showed up when other researchers looked at hypercompetitive individuals. They were “highly narcissistic. . . . At base, however, they were found to have low self-esteem.”25 And it showed up again when yet another group of investigators found that students with a sense of entitlement had lower self-esteem.26

Baumeister himself appeared to back-pedal a few years after publishing his much-quoted “dark side” article. In an essay with a different group of collaborators, he acknowledged that high self-esteem comes in different forms, that many people with high self-esteem aren’t aggressive or narcissistic, and that “psychologists who wish to study or reduce aggression might be well advised to focus on factors other than self-esteem or, at least, to respect the heterogeneity of high self-esteem and therefore consider additional variables.”27 Unfortunately, his first paper had already done its damage: Many people continue to believe there’s something unsavory about having high self-esteem, even though what’s really problematic is (a) low self-esteem, (b) some aspect of self-esteem other than how high or low it is (which I’ll discuss in a moment), or (c) something other than self-esteem.

But don’t many young people have inflated self-esteem?

First of all, let’s keep in mind that, on average, adolescents and young adults do not have higher self-esteem than older adults, nor is there good evidence to support the charge that young people today have higher self-esteem than young people had in years past.28 Most important, though, if there’s really nothing wrong with having high self-esteem, then what exactly is meant by saying that it’s inflated? The answer is that young people feel better about themselves than critics believe they have a right to feel based on what they’ve accomplished.

This determination, of course, is grounded in value judgments, which need to be defended, about what level of achievement “justifies” a given level of self-esteem, and about the underlying belief that the two must be connected at all. But if the charge of inflated self-esteem also has an empirical basis, the claim seems to be that levels of achievement are low (or dropping) while levels of self-esteem are high (or rising). So let’s take a look at that.

Even before the spike in popularity for school-based self-esteem programs, conservatives were already denouncing what they called the “self-esteem movement.” “In the 1950s, before the Self-Esteem-Now theory was widely implemented in American schools,” a writer named Barbara Lerner informed educators in 1985, “competence was widespread, and excellence was common enough to make American students equal to those of any nation. In the 1970s, that was no longer so.” Students in fifth grade and above were “learning less,” she said, at the same time that self-esteem had become “excessive.”29 This story soon became the conventional wisdom, often ungrammatically summarized as “Kids are doing bad and feeling good.”30

The claim that children’s self-esteem has risen over the decades, as I’ve already mentioned, lacks support. The corollary premise—that academic performance has dropped—is persuasive primarily to people whose knowledge of education is limited to what they read in the newspaper, which may help to explain its popularity among politicians. Lerner’s nostalgia is just one more example of the Chicken Little school of thought—“The standards are falling! The standards are falling!”—which, as we saw in Chapter 1, has been on display in every generation.

Those who sound the identical alarms today tend to draw unfavorable comparisons between the current state of our schools and what they were thought to be like a few decades ago. The same is true of self-esteem: Twenge compared students’ noxiously high levels of satisfaction with themselves in 2006 with the more decorous modesty of their counterparts in 1975.31 In both cases, of course, what current curmudgeons see as the good old days (back when achievement was high and self-esteem was low) defined the very period that Lerner described as disgraceful— that is, the 1970s.

But regardless of one’s point of comparison, the idea that students today are “doing bad[ly]” is difficult to support. Even if one looks at the results of standardized tests—a very poor indicator of meaningful learning32—the National Assessment of Educational Progress, commonly described as “the nation’s report card,” fails to support the commonly accepted story of decline. There hasn’t been much change since NAEP testing began, and most of the change that has occurred has been for the better. In both reading and math, when the results for 2012 are compared to those for the early 1970s, some scores were approximately the same and some were markedly higher, depending on which age group and subject you look at.33 But when critics who pound the lectern about declining scores are presented with statistics that show our schools are actually doing about as well as ever, they don’t miss a beat. “Well, doing as well as we used to do just isn’t good enough!” they declare—presumably in an attempt to distract us from the fact that their original claims were baseless.

Either that or they change the subject, switching from a comparison of now to then, to a comparison of us with them—“them” being students in other countries. But here, too, it turns out, the generally accepted story about poor performance on the part of US students is largely incorrect: They’ve held their own overall,34 requiring critics to cherry-pick results to make the situation appear dire. In any case, it’s frankly ridiculous to offer a summary statistic for all children at a given grade level in light of the enormous variation in scores within this country. Test results are largely a function of socioeconomic status. Our wealthier students perform very well when compared to those in other countries; our poorer students do not. And the United States has a lot more child poverty than other industrialized nations.

American kids are sometimes compared to those in Asian countries, a particularly popular pastime after one survey showed that Asian students were critical of their own performance. This was supposed to represent the ideal, and the opposite of our situation: They do well but feel badly anyway: the traditionalist’s dream come true. Less commonly discussed is the fact that many Asian leaders have become increasingly concerned about how their young people may be better at taking tests than at thinking, how they lack the impulse for creativity that’s needed in their society.35 As for that survey—now twenty-five years old—that compared math performance to kids’ beliefs about their aptitude: It may say less about Americans’ swelled heads than about the widely recognized Asian aversion to self-commendation.

But let’s put all of these findings in perspective. Some of the talk about inflated self-esteem doesn’t seem to be based on how well kids are actually doing. Rather, it reflects a way of looking at the world in which a sharp dichotomy is drawn between helping students to feel better about themselves, on the one hand, and spending time on academics, on the other. The former is depicted as a touchy-feely fad, the latter as old-fashioned honest toil. The former amounts to coddling students by pretending everything they do is fine, while the latter means facing up to hard truths and insisting that students measure up to tough standards. It’s almost as if, regardless of the actual effects, attending to self-esteem is regarded as objectionable, a distraction from the unpleasant tasks that must be done, tasks that kids used to do uncomplainingly in the old days before we started worrying about how they felt.

In case it needs to be said, achievement is not undermined by paying attention to how kids feel about themselves—or about their teachers, the curriculum, and the whole experience of school. In fact, these things are closely related. Academic excellence is more likely to flourish when children enjoy what they’re doing and when they do it in an atmosphere of exuberant discovery, in the kind of place where they plunge into their projects and can’t wait to pick up where they left off yesterday. Success is often the result of kids’ feeling confident about—yes, even pleased with—themselves for having figured things out.36 No wonder the statistics belie the claim that a concern about children’s feelings has displaced academic performance. Logic, meanwhile, challenges the charge that self-esteem is inflated.


Peel back the criticism of self-esteem and you’ll eventually uncover the same three values (discussed in chapters 4 and 5) that fuel the attack on overparenting and indulgence. Part of the objection has to do with deprivation: Kids simply shouldn’t be too pleased with themselves. Part of it is based on scarcity: Everyone can’t be a winner, therefore everyone shouldn’t feel good about him- or herself. And the primary dynamic at work here, the one I want to explore at length, is conditionality. To talk about “inflated” self-esteem is to assume that satisfaction with oneself must be justified by a certain level of accomplishment. Kids must earn the right to be happy with themselves by achieving something impressive. What really drives traditionalists to distraction isn’t self-esteem itself or high self-esteem or even programs to raise self-esteem. It’s the idea of unconditional self-esteem: allowing children to feel good about themselves “for no reason,” as Twenge put it, or, in another writer’s phrase, “just for being.”37 To these critics it seems self-evident that, just as praise ought to be offered only on a conditional basis,38 so self-esteem should be thought of “as a reward rather than an entitlement” (Baumeister).39 As important as it is to set stringent conditions for receiving stickers, A’s, or trophies, it’s even more vital to make sure that children feel valuable only on the basis of what they do, not for who they are.

We can begin, as usual, by asking whether this preference for conditional self-esteem rests on a belief that can be tested empirically. I can think of only one: a core assumption about motivation, which Twenge expressed as follows: “If a child feels great about himself even when he does nothing, why do anything? Self-esteem without basis encourages laziness rather than hard work.” A columnist put it this way: “If people are perfect and lovable just the way they are, why should anyone need to change or strive?”40

If this is really supposed to be psychology (rather than theology), it is very bad psychology. There’s absolutely no evidence to support the depressing premise that for things to get done, we need the anxious energy of perpetual self-doubt, the implication of which seems to be that productivity is inversely related to mental health. In reality, someone who has a core of faith in his or her own efficacy and an underlying conviction that he or she is a good person is no more likely than other people—and probably a good deal less likely—to opt for stagnation. As a rule, it’s hard to stop happy, satisfied people from trying to learn or from trying to do a job of which they can be proud.

Why would anyone think that people who generally feel good about themselves will cease making an effort? If self-esteem must be earned in order to create motivation in people, then the underlying assumption is that “human nature is to do as little as necessary,” as a writer named John Powers succinctly described it in an anti-self-esteem essay called “Feeling Good (for Nothing).”41 Ultimately, this reflects the same bleak view that we encountered in Chapter 4: Without the artificial inducements of carrots and sticks, no one would try, learn, work, achieve, or grow because people respond only to extrinsic motivators. And that includes how you feel about yourself, which can be described as extrinsic even though it’s internal.

A thorough refutation of this prejudice, beyond the brief description of intrinsic motivation that I offered earlier, would require a review of the entire fields of personality theory and the psychology of motivation. Numerous studies have confirmed that children are naturally inclined to try to figure things out, to push themselves to do things just beyond their current level. More broadly, the idea that it’s natural to do as little as possible is a relic of “tension-reduction” or “homeostatic” theories, which hold that organisms always seek a state of rest. These theories are taught only in courses on the history of psychology; few beliefs in the social sciences have been so thoroughly repudiated. The desire to do as little as possible is not “human nature”; it is an aberration, a sign that something is amiss.42

The same people who warn that we won’t bother to do anything if we’re happy with ourselves, who believe that self-esteem should rise and fall with our accomplishments in order to goad us into getting off our butts, also tend to put their faith in the usefulness of screwing up and feeling bad. (Twenge: “Sometimes negative feelings can be a motivator.” Powers: “Failure can be a terrific motivator.”43) Of course it makes sense that these views would go hand-in-hand: Low self-esteem is a “negative feeling” about oneself, perhaps the result of a history of failure. But as I’ve noted, the data don’t support claims for the benefits of failure. In fact, studies show that failure is least productive for people with low self-esteem (see pp. 122–23), which suggests that champions of perseverance and learning from one’s mistakes should be defending rather than deriding the importance of self-esteem.

But of course they don’t. And the feature of the whole “self-esteem movement” that elicits particularly withering criticism is loosened conditionality: the possibility that children may be getting something for nothing, or at least for less effort. “Something” includes that which is tangible (trophies), symbolic (grades), and verbal (praise); now, in the case of self-esteem, it extends to the psychological.

The fact is that no research has ever shown that unconditionality has adverse consequences. It may offend the sensibilities of those with what George Lakoff calls a Strict Father morality, but to the best of my knowledge not a single article or book has ever cited evidence showing it to be disadvantageous in terms of future achievement, psychological health, or anything else. In fact, the research reveals exactly the opposite to be true, particularly where self-esteem is concerned.

To understand this, we need to take a step back. I’ve been arguing that self-esteem is a reasonably important variable and that higher self-esteem is generally better than lower. But over the last few years a number of psychologists have suggested that what matters most about self-esteem isn’t how much of it one has. Just as there are different ways of thinking about motivation (for example, whether it’s intrinsic or extrinsic) instead of seeing it as a single entity that can go up or down, the same is true here. In particular, self-esteem varies in terms of stability. If it’s unstable or fragile, even more than if it’s predictably low, the result may be anger or depression. Conversely, even someone whose self-esteem is generally high may struggle with self-doubt and become defensive if that positive view isn’t secure. As one group of researchers put it, “Individuals with fragile high self-esteem are willing to go to great lengths to defend their positive, yet vulnerable, feelings of self-worth. . . . [They] often overreact to perceived threats to their self-worth by becoming angry and either criticizing or attacking the source of the threat.”44

Interestingly, one indication that someone has unstable self-esteem is “a greater tendency to invoke one’s feelings of self-worth in everyday activities.”45 The more variation there is in how you feel about yourself from one situation to the next, the more likely you are to be preoccupied with the whole issue of self-esteem. People with a steady, even if not completely positive, view of their own competence or value don’t have to spend time thinking about how good they are or deliberately trying to feel better about themselves.

But what makes self-esteem unstable? All of us receive feedback that is sometimes positive and sometime negative; all of us do some things we’re proud of and some things we regret. Yet the stability of one person’s self-esteem may be quite different from someone else’s. What accounts for that? The crucial determinant of stability, it seems, is unconditionality. It’s a solid core of belief in yourself, an abiding sense that you’re competent and worthwhile—even when you screw up or fall short—that creates a more stable form of self-esteem, which, in turn, carries a range of psychological and social advantages.

The flip side is that, where self-esteem is concerned, fragility is really a symptom of conditionality.46 If you think well of yourself only under certain conditions—if you regard self-esteem as something that must always be earned and therefore is forever in doubt—then you’re in for trouble, psychologically speaking. Low self-esteem (“I don’t feel very good about myself”) is bad enough; self-esteem that’s contingent (“I feel good about myself only when . . . ”) is even more worrisome.

Contingent on what, though? Some bases for feeling good about oneself may be worse than others. Jennifer Crocker, a psychologist at Ohio State University, and her colleagues have shown that the prognosis is particularly bad when self-esteem hinges on outdoing others (competitive success), approval by others, physical appearance, or academic achievement.47 Consider the last of those. When children’s self-esteem rises or falls with how well they do at school, achievement can resemble an addiction, “requiring ever greater success to avoid feelings of worthlessness.” And if it looks as though success is unlikely, kids may “disengage from the task, deciding it doesn’t matter, rather than suffer the loss of self-esteem that accompanies failure.”

If I need to do well in order to feel good about myself, that may sometimes light a fire under me so I will avoid embarrassing myself. But if I face something really hard, I’ll create an excuse for not having succeeded (which, as I mentioned on pp. 96-97, is known as self-handicapping). Either way, my real goal isn’t to learn; it’s to rescue my shaky belief in my own competence. If I can do that by figuring out a reason why it makes sense that I fell short, then that’s what I’ll choose. The bottom line is that “contingent self-worth is an ineffective source of motivation.”48 People who don’t think their value hinges on their performance, meanwhile, are more likely to see failure as just a temporary set-back, a problem to be solved.

The danger here isn’t limited to achievement-based self-esteem, however. In fact, it may not matter what exactly one has to do or be in order to feel good about oneself. The problem is inherent to the idea of conditionality, which is “associated with anxiety, hostility, defensiveness, and the risk of depression.”49 And that’s not all. Studies have found that contingent self-esteem is related to feelings of helplessness and also to “maladaptive perfectionism.” Teenagers whose acceptance of themselves depends on external factors suffer more intensely if they’re bullied. College students with conditional self-esteem are more likely to drink “as a means of gaining social approval or avoiding social rejection.”50 Some psychologists have suggested that it can also contribute to narcissism and materialistic values. And when people with conditional self-esteem have children of their own, they may base how they feel about themselves on how successful those children are. That not only impairs their own emotional functioning but also leads to a controlling style of parenting. Which may start the cycle all over again in the next generation.51

Even before researchers began collecting all these data, many psychotherapists had figured it out. Alice Miller wrote that one is free from depression “only when self-esteem is based on the authenticity of one’s own feelings and not on the possession of certain qualities.”52 Carl Rogers traced any number of problems to our having been taught to place “conditions of worth” on ourselves; psychotherapists attempt to supply the “unconditional positive regard” that we may have been missing. Experimental evidence seems to confirm that people who have that unconditional faith in themselves are less likely to be anxious or depressed.53

This is powerful stuff. The very unconditionality that is ridiculed by conservatives—indeed, that seems to account for their penchant for vilifying the whole “self-esteem movement”—turns out to be a defining feature of psychological health! It’s precisely what we should be helping our children to acquire. That doesn’t mean harboring unrealistic or grandiose beliefs about one’s own competence. Quite the contrary: It means being sufficiently secure about oneself so as to be able to acknowledge one’s failings and try to improve.

Neither does unconditional or stable self-esteem imply that our feelings about ourselves never vary. It’s natural to be happier (in general, and with oneself) when we’re successful and to be disappointed when we aren’t. But underlying those temporary fluctuations is a permanent reservoir of respect for oneself, a fundamental acceptance of one’s own worth.54 In one of my favorite movies, Harold and Maude, the almost-eighty-year-old Maude admits that, even though she’s impatient with people who are too attached to their possessions, she, too, enjoys “collecting things.” But she immediately qualifies this by describing those things as “incidental, not integral, if you know what I mean.” That’s an apt way of thinking about our successes and failures with respect to self-esteem: Their impact isn’t integral to how we see ourselves.


So what determines whether people place conditions on the way they regard themselves? There are no hard-and-fast answers here; no one has ever conducted a study that can settle the matter once and for all. In fact, there may not be a single answer for everyone. But the best guess is that conditional self-esteem results from having been esteemed conditionally by others. When children feel as though they must fulfill certain conditions to be loved by their parents—a feeling typically evoked by the use of psychological control—it’s not easy for them to accept themselves unconditionally. And everything goes downhill from there.

Susan Harter, a developmental psychologist at the University of Denver, points out that “level of support and conditionality of support are correlated with one another.” In other words, the kind of parents who are not especially warm and loving also tend to be the kind who put conditions on the affection they do provide. But even when you hold the level of support constant, the degree of conditionality counts. “At relatively high as well as relatively low levels of support,” she explains, “the more conditional the support, the lower one’s self-esteem55—and the more likely it is that one’s self-esteem will also be conditional.56

That’s not really surprising, is it? Children don’t just need to be loved; they need to know that nothing they do will change the fact that they’re loved. They require reassurance that their “lovability” isn’t in question, which is another way of talking about self-esteem. By contrast, one conservative critic of self-esteem not only complains about “unearned praise” for children but expresses distaste for how “today’s parents” are likely to express “enthusiasm for their children’s very existence.”57

Would that it were true! The eminent psychoanalyst and social critic Erich Fromm put it this way:

Unconditional love corresponds to one of the deepest longings, not only of the child, but of every human being; on the other hand, to be loved because of one’s merit, because one deserves it, always leaves doubt; maybe I did not please the person I want to love me, maybe this, or that—there is always a fear that love could disappear. Furthermore, “deserved” love easily leaves a bitter feeling that one is not loved for oneself, that one is loved only because one pleases, that one is, in the last analysis, not loved at all but used.58

Beginning in 2004, two Israeli researchers, Avi Assor and Guy Roth, and their colleagues have been conducting experiments to examine this issue scientifically. They started by asking college students whether the love they had received from their parents seemed to depend on whether they had succeeded in school, practiced hard for sports, been considerate toward others, or suppressed emotions such as anger and fear. The answer: Yes, that was indeed true for many of them. (Their parents probably would have been dismayed to hear they felt that way. But what determines a psychological outcome isn’t what parents think they’re doing; it’s how the children experience it.) It turned out that kids who received what they interpreted as conditional approval were indeed somewhat more likely to act the way the parent wanted. But that compliance came at a steep price. First, these children tended to resent and dislike their parents. Second, they were more likely to attribute their behavior to “strong internal pressure” than to “a real sense of choice.” Moreover, their satisfaction after succeeding at something was usually short-lived, and they often felt guilty or ashamed.

The researchers then interviewed a group of mothers of grown children. With this generation, too, conditional parenting had proved damaging. Those mothers who, when they were young, sensed they were loved only when they lived up to their parents’ expectations now felt less worthy as adults. Yet despite the negative effects, these mothers were more likely to use conditional affection with their own children.59 Apart from being profoundly depressing, that finding can serve as a metaphorical Rorschach test, revealing something about one’s way of thinking about such issues. Is it mysterious and counterintuitive? (Why on earth would someone do to her children exactly what damaged her?) Or is it perfectly predictable? (How could someone love her children unconditionally if she never received that kind of love herself?)

Assor, Roth, and their collaborators have gone on to publish a variety of replications and extensions of these original studies. In one of them, their subjects were ninth graders, and this time giving more attention and affection when children did what parents wanted was carefully distinguished from giving less when they didn’t. The studies showed that both positive and negative versions of conditional parenting were harmful, but in slightly different ways. The positive kind—praise for success—sometimes succeeded in getting children to work harder on academic tasks, but, again, at the cost of unhealthy feelings of “internal compulsion.” Negative conditional parenting, meanwhile, didn’t even work in the short run; it just increased the teenagers’ resentment toward their parents.60

In their other studies, which were conducted with children from age five to twenty-three, conditional parenting had consistently disturbing effects on their emotional and social well-being:

• If young children got the idea that their parents valued them more when they were happy, that interfered with their ability to recognize and respond to sadness in other people.

• If young adults had perceived that their parents’ affection varied with the extent to which they were helpful and considerate, that affected the way these grown children thought about helping: It seemed less a matter of choice than something they felt they had to do to try to feel better about themselves.

• If teenagers got the idea that their parents’ approval of them depended on how well they did in school, they were apt to be self-aggrandizing when they succeeded and ashamed when they failed. “Conditional positive regard promote[d] the development of a fragile, contingent . . . and unstable sense of self.”61

Regardless of the child’s age, regardless of what behavior is required as a condition of the parent’s love, and regardless of whether love is offered for engaging in that behavior or withheld for not engaging in it, the outcomes are troubling. (In fact, they’re troubling even when a teacher rather than a parent seems to accept children only conditionally.62) One way of making sense of these findings, I’ve argued, is to consider the creation of conditional self-acceptance. The adult’s “My care for you depends on your doing x” becomes the child’s “I’m worthwhile only if I do x.” Research confirms that conditionality is a recipe for dysfunction.

But our objections may go beyond what the data say. In a scholarly article entitled “Contingencies of Self-Worth,” Jennifer Crocker and Connie Wolfe took two dozen pages to review empirical findings on the topic. Then, at the very end, they took the unusual step of adding a personal note:

We are alarmed at the suggestion that schools should be teaching or creating self-esteem that is justified by achievements or “warranted” (Baumeister, 1999; Seligman, 1998). Such recommendations are equally lacking in empirical support. Furthermore, and perhaps more seriously, the logical implication of this approach is that some children’s low self-esteem is warranted and that children who do not achieve in socially desirable ways, such as getting good grades, being attractive and popular, or being good at sports, rightly believe that they are not worthy human beings. It is a slippery slope from the view that self-esteem can be warranted or unwarranted to the view that some people are unworthy and justifiably devalued.63

The attack on “empty” or “excessive” praise, or on “unearned” grades or trophies, is not just about what adults offer to kids; it’s ultimately about how kids view themselves. There’s no evidence to support the idea that self-esteem is “inflated” or that children do better when their view of themselves varies with their performance (or with anything else). But once again, this critique is based less on evidence than on the value judgment that feeling good about oneself is something one ought to have to earn. Crocker and Wolfe confronted that value judgment head-on, saying, in effect, that this is an appalling way to raise and educate children. How well you do things should be incidental, not integral, to the way you regard yourself.



[For full citations, please see the Reference section of The Myth of the Spoiled Child.]

1. Mecca.

2. For example: Christopher Lasch (“For Shame: Why Americans Should Be Wary of Self-Esteem”) in 1992, Joseph Adelson (“Down with Self-Esteem”) in 1996, Chester E. Finn, Jr. (“Narcissus Goes to School”) in 1990, Al Shanker (“All Smiles”) in 1994, and Charles Krauthammer (“Education: Doing Bad and Feeling Good”) in 1990.

3. Of course, some people do compliment kids to help them feel better about themselves. But not all self-esteem boosting consists of praise, and, more important, not all praise is intended to raise self-esteem. As I argued earlier, much of it, like other rewards, is more about reinforcing certain behaviors in an effort to elicit compliance.

4. For example, Martin Seligman’s dismissive comments about self-esteem appear alongside his denunciation of “massive grade inflation . . . competition becoming a dirty word . . . the demise of rote memorization . . . less plain old hard work” (Seligman, p. 28). And Roy Baumeister, the most prominent academic critic of self-esteem, advises a sympathetic interviewer (in an article titled “When Bad Kids Think They’re Great”) to “set the rules, reward the child when the child does well, punish the child when he or she does badly” (Milstone).

5. “Poorer mental and physical health . . . ”: Trzesniewski et al. 2006. Preadolescents more aggressive: Donnellan et al. Problem eating: McGee and Williams. Depression: Orth et al. 2012: “We replicated previous studies showing that low self-esteem prospectively predicts depression but that the effect of depression on low self-esteem is small or nonsignificant. . . . A similar pattern emerged for measures of dispositional positive and negative affect. . . . In addition, we found that self-esteem was prospectively related to higher levels of relationship satisfaction, job satisfaction, occupational status, salary, and physical health, controlling for prior levels of these variables, but none of these life outcomes had reciprocal effects on self-esteem (or, if significant, the coefficients were small). Moreover, all results held across generations” (p. 1283).

6. Baumeister et al. 2003, pp. 25, 28; Lyubomirsky et al.

7. Baumeister et al. 2003, p. 15; McFarlin and Blascovich.

8. Swann et al., p. 87.

9. O’Mara, p. 184. Also see Valentine et al.

10. Baumeister et al. 2003, p. 14. Oddly, Baumeister continues to tell interviewers that “self-esteem is a result, not a cause” (Stein, p. 28), even though his own review of the research challenges this simplistic statement. Twenge (2006, p. 65), too, offers the unsupported view that “self-esteem does not cause high grades—instead, high grades cause higher self-esteem.”

11. One group of researchers argued in the early 1980s that two variables— social class and academic ability—account for most of the variance in self-esteem and in performance. See Maruyama et al.

12. DuBois and Tevendale, p. 110. Of course, plenty of other things, too, have an impact on both self-esteem and achievement.

13. O’Mara et al., p. 185.

14. See Kohn 1992, esp. chapter 5.

15. Children do not come to believe they are important, valued, and capable just because they are told this or made to recite it. But programs that are ineffective and, in some cases, verge on self-parody are sometimes used to discredit the whole enterprise of enhancing kids’ self-esteem—indeed, the very idea of self-esteem. And some traditionalists don’t stop there: They dismiss any school-based efforts to deal with children’s emotional lives or address anything beyond narrowly defined academic skills. The case for attending to students’ social and emotional needs at school is overwhelming, even if one is interested only in academic outcomes. For three of many examples, see Flook et al.; Reyes et al.; and Zins et al.

16. First review: Haney and Durlak (quotation on p. 430). Second review: O’Mara et al. (quotation on p. 195).

17. Baumeister has also argued that because (a) men commit more crimes than women and (b) men tend to score higher on measures of self-esteem (“although the difference is not large and may be diminishing in the modern world”), it follows that there’s something undesirable about high self-esteem (Baumeister et al. 1996, p. 13). Even more outrageously: Black men now rape white women more often than white men rape black women—a shift that coincided with “efforts to boost self-esteem among Blacks” (ibid., p. 14).

18. Twenge 2006, p. 224. Here she assumes that making demands, which of course is just part of what it means to be a young child, reflects a particular view of the self. She also applies the concept of self-esteem (which was developed with reference to adults and older children) to toddlers, whose basic sense of self isn’t even fully formed. And she continues, “Even as children grow older, most are confident and self-assured”—a statement that could charitably be described as bizarre.

19. Otway and Vignoles, p. 104.

20. Gilligan quoted in Weissbourd, p. 13. interestingly, Baumeister seems to accept another writer’s proposition that “to avoid certain negative emotional states, such as shame, dejection, sadness, and disappointment with oneself, the [aggressive] person refuses to contemplate information that reflects unfavorably about the self” (p. 11). What he never explains is how that profile is consistent with having high self-esteem. At another point, he insists that “only the person with a highly favorable opinion of self will be inclined to seek out risky situations to prove his or her merit” (p. 7). But why would such a person need to do so? Those who are easily threatened and inclined to lash out are more likely to hold a tenuous opinion of themselves.

21. Rosenberg quoted in Donnellan et al., p. 333.

22. See p. 201n57.

23. Tracy et al., p. 209.

24. “A robust relation between low self-esteem and externalizing problems . . .held for different age groups, different nationalities, and multiple methods of assessing self-esteem and externalizing problems” (Donnellan et al.; quotation on p. 333).

25. Ryckman et al.; quotation on p. 91. This replicates another study published four years earlier by Ryckman and a different group of collaborators.

26. Greenberger et al. Three years later, these researchers looked separately at exploitive and non-exploitive entitlement. Self-esteem was negatively correlated with the former and positively correlated with the latter. See Lessard et al.

27. Baumeister et al. 2003, p. 24. Also see pp. 5–6, 21.

28. Self-esteem tends to increase through adulthood until about age sixty (see p. 201n58). As for now-versus-then comparisons, see the research findings from several groups of scholars that contradict Twenge’s claims, which I described on pp. 28–29.

29. Lerner, p. 15.

30. Krauthammer.

31. Twenge and Campbell 2008.

32. See Kohn 2000, which, in turn, cites many other sources.

33. You can check this yourself at Also see Berliner and Biddle; Rothstein; and many writings by the late Gerald Bracey.

34. Kohn 2013.

35. For two of many examples, see Zhao; and the comments of Byong Man Ahn, South Korea’s former minister of education, science, and technology, quoted in Cavanagh.

36. It’s important to add that self-esteem isn’t important only for its contribution to academic performance. Quality of life and joy in learning are ends in themselves.

37. Twenge 2006, p. 57; Ehrensaft, p. 123.

38. See pp. 216–17n11.

39. Baumeister et al. 2003, p. 39.

40. Twenge: 2006, p. 66. The columnist: Leo. (Note how he conflates being lovable with being perfect.)

41. Powers, p. 8.

42. Interested readers might look up the work of Gordon Allport as well as findings concerned with the fundamental human impetus to attain a sense of competence (Robert White), to be self-determining (Richard deCharms, Edward Deci, and others), to satisfy our curiosity (D.E. Berlyne), and to “actualize” our potential in various ways (Abraham Maslow).

43. Twenge 2006, p. 66; Powers, p. 8.

44. Kernis et al. 2008, pp. 478, 479. For more about the importance of stability or fragility in evaluating self-esteem, also see Kernis; Kernis et al. 1993; and Seery et al.

45. Kernis et al. 2000, p. 245.

46. On this point, see Tracy et al., especially p. 4; and Deci and Ryan 1995. One difference between the two syndromes, though, is that people with unstable self-esteem may not be aware of that fact, whereas people with contingent self-esteem know what has to happen in order for them to feel good about themselves (see Kernis et al. 2008, p. 501).

47. For example, see Crocker and Knight; and Crocker and Wolfe.

48. Crocker and Knight, pp. 200, 202.

49. Crocker and Wolfe cite eight studies to substantiate that first list of unhappy consequences, from anxiety to depression (p. 614). Since then, Burwell and Shirk also found that contingent self-esteem, even more than low self-esteem, is a risk factor for depression in adolescents.

50. Helplessness: Burhans and Dweck. (The finding that “children who had expressed a sense of contingent self-worth were significantly more helpless” was replicated in a subsequent study by Kamins and Dweck, described in Dweck, p. 115.) “Maladaptive perfectionism”: Soenens and Vansteenkiste. Impact of bullying: Ghoul et al. Drinking: Neighbors et al.

51. Narcissism: Assor and Tal, p. 257. Materialism: Ku et al., pp. 83–84. Effects on parenting: Eaton and Pomerantz; Grolnick et al.; and Ng et al.

52. Miller, p. 58.

53. Chamberlain and Haaga.

54. Also, the reactions of “pleasure following success and disappointment following failure . . . are not colored with defensiveness or self-aggrandizement” in people with stable, unconditional self-esteem (Kernis et al. 2008, p. 500).

55. Harter, p. 101. Emphasis added.

56. See Wuyts et al.

57. Young-Eisendrath, p. 27.

58. Fromm, pp. 41–42.

59. Assor et al. 2004.

60. Roth et al.

61. Kindergarteners: Roth and Assor. Young adults: Roth. Teenagers: Assor and Tal. (Quotation on p. 257.)

62. In a study of three hundred middle schoolers, students had lower self-esteem and less intrinsic motivation to learn if they felt their teachers’ acceptance of them was contingent on their achievement or on having met the teachers’ expectations (Makri-Botsari).

63. Crocker and Wolfe, p. 617.

To be notified whenever a new article or blog is posted on this site, please enter your e-mail address at .