Self-Esteem and Student Achievement

develop through effort and education.They don't necessarily believe thatanyone can become an Einstein or aMozart, but they do understand thateven Einstein and Mozart had to put inyears of effort to become who they were.When students believe that they candevelop their intelligence, they focus ondoing just that. Not worrying abouthow smart they will appear, they takeon challenges and stick to them(Dweck, 1999, 2006).More and more research inpsychology and neuroscience supportsthe growth mind-set. We are discoveringthat the brain has more plasticityover time than we ever imagined(Doidge, 2007); that fundamentalaspects of intelligence can be enhancedthrough learning (Sternberg, 2005);and that dedication and persistence inthe face of obstacles are key ingredientsin outstanding achievement (Ericsson,Charness, Feko\'ich, & Hoffman,2006).Alfred Binet (1909/1973), theinventor of the IQ test, had a stronggrowth mind-set. He believed thateducation could transform the basiccapacity to learn. Far from intending tomeasure fixed intelligence, he meant histest to be a tool for identifying studentswho were not profiting from the publicschool curriculum so that other coursesof study could be devised to foster theirintellectual growth.The Two Faces of EffortThe fixed and growth mmd-sets createtwo different psychological worlds. Inthe fixed mind-set, students care firstand foremost about how they'll bejudged: smart or not smart. Repeatedly,students with this mind-set rejectopportunities to learn if they mightmake mistakes (Hong, Chiu, Dweck,Lin, &r Wan, 1999; Mueller & Dweck,1998). When they do make mistakes orreveal deficiencies, rather than correctthem, they try to hide them (Nussbaum& Dweek, 2007).They are also afraid of effort becauseeffort makes them feel dumb. Theybelieve that if you have tbe ability, youshouldn't need effort (Blackwell, Trzes-Promises of Praise

Min dsets:How Praise bs Harming Youthand What Can Be Done about Itby Carol S. DweckToday's companies arereporting that theiryoung employees needconstant praise. Withoutit, they become moroseand disgruntled. Coachesare complaining that theirathletes can no longer toleratecorrective feedback. Theathletes claim it underminestheir confidence. Parents saythat their children don't wantto work hard in school. Thechildren feel it should justcome naturally.What has happened in this countrythat used to be known for a solidwork ethic and rugged individualism?Now, young people have a disastrouscombination of entitlementand fragility. How did this happenand what can be done about it?Research conducted at Columbiaand Stanford Universities suggeststhat it could be well-meaning parentand teacher practices that havebrought this on-practices that weremeant to boost childrenfs self-esteem.In the 1990s, parents and teachers,prompted by the self-esteem movement,decided that self-esteem wasthe most important thing in theworld-that if a child had self-esteemeverything else would follow.The big problem came when theydecided how to give children thatself-esteem. In many cases, parentsand teachers begantelling children ona regular basis howsmart and talentedthey were. In a surveydone in the mid1990s, over 80% ofparents reportedthat they thoughttheir children must 'be praised in orderto give them confi- --..dence and motivation(Mueller andDweck 1998). Mostof them still seem tobelieve this. Recently, in an airport,the author observed a mother tellingher 6-month-old son that he wasbrilliant.Unfortunately, this commonlyheld belief of both teachers and parentsisn't true. Research shows thatchildren cannot be given self-esteemthrough this kind of praise (MuellerCarol S. Dweck is the Lewis and Virginia Eaton Professor of Psychology at StanfordUniversity, Stanford, CA, and the author of Mindset (Random House, 2006).and Dweck 1998; Kamins and Dweck1999). But, even worse, it also showsthat this type of praise actually makestheir self-esteem fragile-and underminestheir motivation to learn. Howcan this be? How can the effect bejust the opposite of what intuition orcommon sense seem to indicate?Understanding MindsetsThe mindsets of students mustfirst be understood to appreciate whypraise for intelligence or talent mightbackfire. For over 20 years researchhas explored these mindsets andshown their impact on students' confidenceand motivation (Dweck 2006).Research indicates that some studentshave a fixed mindset becausethey believe that their intelligenceSchoolLibraryMediaActivitiesMonthly/Volume XXIV, Number 5/January200855

is simply a fixed trait. They have acertain amount, and there's nothingthey can do to change it. Not surprisingly,they worry about how muchthey have. Other students, however,have a growth mindset. They believethat their intelligence is a quality thatcan be developed. They don't believeeveryone is the same, but they do believethat everyone can increase theirintelligence through effort and education.Therefore, they don't spendtheir time worrying about how smartthey are and instead spend it tryingto get smarter.Which view is correct? Althoughthe answer to such questions istypically that both views have merit,more and more research indicates thefollowing:"• Important aspects of intelligencecan be developed (Sternberg2005)."* The brain has enormous capacityto grow and change throughoutlife (Doidge 2007)."• Motivation and self-regulationoften have more impact onachievement and the growth ofability than does a person's initialability (Blackwell et al. 2007;Duckworth and Seligman 2005;Ericsson et al. 2006).These differing mindsets affectstudents' achievement by creatingdifferent psychological worlds. Hereis how these worlds work.What do students care about?In the fixed-mindset world, studentscare more about looking smartthan about learning. In their world,every performance holds their intelligenceup for judgment, so thatlearning takes a back seat to lookingsmart. In fact, research by Dweckand her colleagues has shown thatstudents in a fixed mindset will evensacrifice learning that is crucial totheir future success if they have toadmit ignorance or risk showing deficiencies(Hong et al. 1999).On the other hand, in the growthmindsetworld, students care firstand foremost about learning. Theydon't have to worry about discreditingtheir permanent intelligence, sothey can devote themselves to thetask of getting smarter.What do mistakes and effortmean?In the fixed-mindset world, studentsworry about making mistakes.They see making mistakes as asign of low ability. They also worryabout effort and view it in the sameway-as a sign of low ability. Theybelieve that if they have high abilitythey shouldn't need any effort. Theseare both terrible beliefs because mistakesand effort are integral parts oflearning. Because students with fixedmindsets make them into things tobe avoided, they actually stand in theway of learning. In fact, research inpsychology indicates that the mainthing that distinguishes people whogo to the top of their fields and makegreat creative contribution from theirequally able peers is the effort theyput in (Ericsson et al. 2006). Thefixed mindset cannot take people tothat level.Students with a growth mindsetunderstand that mistakes and effortare critical to learning. They, therefore,welcome challenges and seekcritical feedback to help them learn.When students are asked, "Whendo you feel smart?' students witha fixed mindset reply that they feelsmart when something is easy forthem, when they do a task quicklywithout mistakes, or when otherpeople look dumb..But students witha growth mindset state that they feelsmart when they are struggling withsomething hard and make progressor when they are helping someoneelse to learn (Dweck 2006).Thus, it is the growth mindset thathelps students adopt the values thatlead to intellectual growth.How do students react tosetbacks?Students with a fixed mindsetlose heart in the face of setbacks. Ifsetbacks mean they lack ability andability is fixed-and effort is distasteful-whatcan they do? Compared tostudents with a growth mindset, studentswith a fixed mindset indicatethat after one failure in a class theywould withdraw their effort, avoidsimilar courses in the future, andconsider cheating. The fixed mindsetprovides no good recipe for recoveringfrom a failure.Students with a growth mindsetsay that after a failure they wouldsimply study more in the future orstudy in a different way. Their failureis a spur to learning. As a result,these students end up earning highergrades-whether it's during theirtransition to junior high or the challengingorganic chemistry course atthe beginning of the pre-med collegecurriculum.In short, students with a fixedmindset, no matter how bright, oftendevelop values and habits that standin the way of developing their abilitiesand doing well in school. In contrast,students with a growth mindsetembrace learning, mistakes, andeffort in a way that promotes theirachievement.Praise: How Do StudentsLearn the Mindsets?What effects do praise and otherwell-meaning practices of parentshave on their children? Modern parentswant to give their children everythingpossible to build them up andensure their happiness and success.The self-esteem movement told themthat the secret lay in praise, especiallyin praise of the child's essential qualities.This has great intuitive appeal-56 SchoolLibraryMediaActivities Monthly/VolumeXZIV, Number 5/January 2008

tell children they're smart and they'llhave faith in their intelligence.However, studies conducted overmany years involving vulnerable studentsindicated that it was these verychildren who were overly concernedwith how smart they were (Dweck2006). It appeared that parents whoplace too much emphasis on intelligenceor talent could send the wrongmessage. By praising a child's intelligenceor talent, they could be conveyingto their child that intelligenceor talent is something deep and permanentthat can be judged and quantified-afixed mindset. They couldalso be conveying that brains and talentare what they value the child for,so that children become afraid that ifthey're not successful, they won't beconsidered smart anymore and theywon't be valued anymore.Research put this issue to the test(Mueller and Dweck 1998). Severalhundred children-from differentregions of the country and of differentethnicities-were given problemsfrom an IQ test. After the first set ofproblems, some of them were givenpraise for their intelligence: "That's areally good score. You must be smartat this' Others were given praise fortheir efforts: "That's a really goodscore. You must have worked reallyhard' This one line or praise had acascade of striking effects. First, theintelligence praise instilled more of afixed mindset, making students believethat their intelligence was a fixedtrait, whereas the effort praise instilledmore of a growth mindset. Next,the praise changed students' desirefor challenge and learning. Studentspraised for their intelligence nowwanted easy tasks so they wouldn'tjeopardize their label "smart' Studentspraised for effort wanted challengingtasks they could learn from, even ifthey would make mistakes.When given more difficultproblems, the intelligence-praisedstudents crumbled. They lostconfidence, they lost their enjoymentof the task, and their performance-evenon subsequent easierproblems-fell significantly. In thepublic view, intelligence praise wassupposed to hand them the gift ofconfidence, but instead it made themfragile. They were initially puffed upby the intelligence praise, displayingproud little smiles on their facesthat weren't seen in the effort-praisedgroup. But their pride evaporatedquickly at the first sign of difficulty.Effort-praised students, however,maintained their faith in theirabilities, kept on enjoying the task,and showed markedly improvedperformance over time. The effortpraise kept them focused on whatmattered for their motivation andperformance. Rather than worryabout their intelligence, they focusedon their efforts and kept on striving.This is reminiscent of the childrenwho are puffed up by their parents'praise, but then are unable or unwillingto work hard in school. It is alsoseemingly reflective of the youngworkers who need constant praiseand encouragement to keep up theirconfidence.In the studies of praise, there wasone more finding that was troubling.Students were asked to writeanonymously to a child in anotherschool, and a space was left on thesheet of paper for them to writein their score. Almost 40% of theintelligence-praised students liedabout their score. (Only 13% of thestudents in the effort-praised groupreported an incorrect score.) Thismeans that praising children's intelligencemakes them so invested intheir intelligence that they cannotadmit their mistakes. If they cannotadmit their mistakes, they cannotcorrect their mistakes,In several other studies, studentsin a fixed mindset did not confrontand remedy deficiencies-even whengiven the opportunity-whereas studentsin a growth mindset did (Honget al. 1999; Mangels et al. 2006; Nussbaumand Dweck 2008). It is absolutelycritical for people of all ages tohonestly consider their mistakes andfailings because there will always bemistakes and failings no matter howproficient people become. Withoutthis capacity, intellectual growth isstunted.This may well be why coachesare suddenly finding their youngplayers to be uncoachable. Becausethese players are part of the overpraisedgeneration, they may, likethe intelligence-praised students inthe research studies, have gotten themessage that mistakes are unacceptable.They, therefore, may not be ableto take even constructive criticismfrom their coaches-feedback that isessential for their continued growthand success.How to Promote a GrowthMindsetWhat are the ways in whicheducators and parents can foster agrowth mindset in children?FeedbackFollowing up on the praise research,the first way is through processpraise. That is, adults should praisethe process the child engaged in-theeffort, the strategies, the perseverance,the choices-rather than the intelligenceor talent they think the childshowed. Here are some examples:"I really like the way you did thosehard problems. You stuck tothem until you figured out howto do them:'"Those are really interesting colorsyou chose for your painting. Tellme about them:""You practiced a lot and you reallyimproved. That's exciting!"An important role for parents andSchoolIAbraryMediaActivitiesMonthly/VolumeXXIV, Number 5/January2008 57

AS EDUCATORS, ALMOST EVERYTHING WE SAY TO OURstudents sends a message. Some messages enhance students'motivation, but other messages undermine it. Howcan we know which is which? Common sense and intuitionwill always be a part of good teaching, but they are notalways trustworthy guides. This is where research comesin, and helps us put our common sense to the test.Imagine a brilliant student who enters a new school andsuddenly starts getting poor grades. Or a struggling studentwho needs encouragement. Or a talented child wholacks confidence. What should teachers say to these studentsto send messages that motivate? In each case, teachersmight be tempted to look for opportunities to praisethe students' abilities and assure them of their intelligence.In a survey we gave to parents, over 80% of them thoughtthat it was necessary to praise their children's intelligence inorder to give them confidence in their abilities and motivatethem to succeed.Our research shows that this is wrong. As you read on,you will find out why.WHO ARE THE MOTIVATED CHILDREN?For over 30 years, I have studied students' motivation inorder to find out what makes motivated students tick. Hereis the most important thing I have learned: The most motivatedand resilient students are not the ones who thinkthey have a lot of fixed or innate intelligence. Instead, themost motivated and resilient students are the ones whobelieve that their abilities can be developed through theireffort and learning.Is intelligence something inherent or is it something thatcan be developed? Although this is not really an either/orquestion, more and more research is revealing that importantparts of intelligence can be developed and that thebrain has greater potential for growth and changethroughout life than anyone ever thought. It's also interestingtoknow that Alfred Binet, the man who invented the IQtest, profoundly believed that children's intelligence couldbe transformed and enhanced through education. In fact,he devoted most of his career to developing educationalcurricula that would do just that. (The IQ test was simplycreated to identify children who were not profiting fromthe curriculum in the Paris public schools, so that Binetcould develop courses of study that would better meettheir needs.)My main point will be: It matters greatly what studentsbelieve about their intelligence.In a recently-published study, my colleagues (Lisa Blackwelland Kali Trzeniewski) and I followed over 400 studentsacross the transition to seventh grade, a time when manystudents are derailed. At this time, the work gets harder,the grading becomes more stringent, and the environmentbecomes less personalized and nurturing. Students, as awhole, show a concomitant decline in motivation andgrades. However, this was not equally true for all studentsin our study.It was the students who believed in fixed intelligencewho fared most poorly across this transition - even manywho had done well in the past. They showed poorer motivation,less resilience in the face of difficulty, and lower6 EDUCATION CANADA I CANADIAN EDUCATION ASSOCIATION

EN BREF Bien qu'il solt logique de penser que les 616ves non motiv6s, ondifficult6 ou peu sOrs d'eux pourraient b6n6ficier d'61oges au sujet de leurintelligence, Ia recherche le contredit. Les 616ves qui croient que leur succ6sest attribuable I leur intelligence r6ussissent moins que ceux qui croientqu'ils doivent investir des efforts pour r6ussir. Les 616ves ayant I'attitudeancr6e quo leur intelligence est A Ia source de leur rendement adoptent desstrat6gies pour avoir I'air intelligent - m6me s'ils apprennent alors moins. Parcontre, les 616ves anim6s d'une attitude o de croissanceocrolent qu'ils peuventd6velopper leurs habilet6s et adoptent des strat6gies rehaussant leurshabitudes de travail et accroissant leurs apprentissages.grades over the next two years. Those who believed theirintelligence could be developed showed increasing gradesover the same period. How did this happen?THE FIXED MINDSETLet's look first at the students who believed in fixed intelligenceand see how this fixed mindset worked to limit theirachievement. The fixed mindset comes with "rules," thecardinal rule being: Look smart at all costs. Not surprisingly,this rule stands in the way of learning.For example, when our seventh graders were given achoice between learning something new and doing a taskthat would make them look smart, they chose the latter. Inanother study, we polled new students at the University ofHong Kong, an elite university where all classes are in English.We asked students who had poor English skillswhether they would take a remedial English course if thefaculty offered it. Students with a fixed mindset were notenthusiastic - they didn't want to be in a situation wherethey would not look smart. Better to put their collegecareer in jeopardy! (Students with a growth mindset - whobelieved their intelligence could be developed - were farmore enthusiastic about the course.)Of course, wanting to look smart can sometimes bemotivating, but the fixed mindset often comes with threemore rules.1. Don't make mistakes. Students in a fixed mindset t nk`that mistakes or setbacks mean they lack ability. Our seventhgraders told us that if they got one poor grade in anew course, it would mean they weren't good at that subject.They went on to say that they would try to drop thecourse and never take that subject again. In a study withpre-med college students at Columbia University we foundthe same thing. Students in a fixed mindset thought onedisappointing grade measured their ability and their performancenever recovered.In short, students in a fixed mindset believe that if theyhad the intelligence, it would carry them straight throughto perfect performance. Anything less spells inadequacy.This is why many talented students lack confidence inthemselves. Which brings us to the next rule.2. Don't work hard. Our seventh graders in a fixed mindsetthought that hard work signaled low intelligence: "Totell the truth, when I work hard at my schoolwork it makesme feel like I'm not very smart." It didn't matter whetherthe schoolwork was new or difficult; their effort was a signof limited ability.THE IDEA THAT HIGH EFFORT EQUALS LOW ABILITY IS ONE OF THEWORST BELIEFS STUDENTS CAN HAVE. IT IS VIRTUALLY IMPOSSIBLETO DO ANYTHING WORTHWHILE WITHOUT SUSTAINED EFFORTCANADIAN EDUCATION ASSOCIATION I EDUCATION CANADA 7

This is precisely why many "gifted" students simply stopworking when school becomes more difficult. Before, they'dbeen able to coast along and do extremely well. In a fixedmindset, that's how they knew they were gifted. Suddenly,they can't do that anymore. Now they have a choice: workhard and feel dumb (and worst of all maybe still not do spectacularly)or don't work hard and act smart. Act as thoughyou could do well if you wanted to. You just don't care to.The idea that high effort equals low ability is one of theworst beliefs students can have. It is virtually impossible todo anything worthwhile without sustained effort. This iswhy we should not tell students how great they are whenthey do something quickly and easily. They should notthink that low effort is the hallmark of intelligence.3. If you make mistakes, don't try to repair them. Ourseventh graders with a fixed mindsettold us that if they didpoorly on a test, they would study less the next time andseriously consider cheating. Hardly a recipe for success!Actually, the fixed mindset does not provide good recipesfor recovering from setbacks. Setbacks indicate a lack ofability and, in the fixed mindset, that lack of ability is permanent.In one study, we monitored students' brain waves (EEGs)as they performed a very difficult task. What were studentspaying most attention to? The brain waves revealedthat students in a fixed mindset were vitally interested inwhether they got an answer right or wrong, but, when theywere wrong, they paid little attention to what the rightanswer was. They were not trying to correct their errorsand, as a result, did significantly worse than students witha growth mindset when they were later retested on thematerial.I think we can begin to understand how a fixed mindsetcan limit students' learning. Look smart at all costs. Don'tmake mistakes. Don't work hard. If you make mistakes,don't try to correct them. Clearly, these are not rules thatfoster intellectual growth.THE GROWTH MINDSETStudents with a growth mindset believe that their abilitiescan be developed, and so their major goal is to learn. "It'smuch more important for me to learn things in my classesthan it is to get the best grades." Although these studentscare very much about doing well in school, they put a premiumon learning. Ironically, this leads them to earn highergrades.In other words, the cardinal rule of the growth mindset is:Learn! And like the fixed mindset, the growth mindset comeswith three more rules that help students reach their goal.1. Take on challenges. We've often offered students achoice between a challenging task that they can learn fromand a task that is sure to make them look smart. Studentsin a growth mindset do not want to waste their time lookingsmart on tasks that offer them nothing else. They overwhelminglywant tasks that stretch their abilities and teachthem new things.2. Work hard. Rather than thinking that effort underminesability, our seventh graders with a growth mindsetbelieved that effort enhanced ability: "The harder you workat something, the better you'll be at it." They did not believethat inherent ability was the royal road to success, for evengeniuses, they correctly believed, had to work hard for theirsuccesses.3. Confront your deficiencies and correct them. In justabout every study we've done, students in a growth mindsetare eager to remedy their deficiencies. They may be verydisappointed by a poor performance, but they deal with itdirectly. Our seventh graders with a growth mindset, aftera poor grade on a test, told us that they would study harderand try different study strategies next time. Our premedcollege students with a growth mindset recoveredfrom an initially disappointing grade through their deeperand more extensive study strategies. Our University ofHong Kong freshmen wanted the remedial English courseto shore up their language skills. And the growth-mindsetstudents in our EEG experiment confronted their mistakesand sought new knowledge to rectify them.Clearly these are rules and practices that promote intellectualgrowth. What can educators do to foster a growthmindset in their students?MINDSET MESSAGES: PRAISETo answer this question, let's return to the issue of praiseand the message it sends. Although common sense maysuggest that students who are unmotivated, struggling, orlacking in confidence might benefit from praise for their8 EDUCATION CANADA I CANADIAN EDUCATION ASSOCIATION

intelligence, we worried that this kind of praise might senda fixed mindset message. We worried that praising students'intelligence, even after a job well done, might tellthem: 1) I can look at your performance and judge yourunderlying intelligence (a fixed mindset message) and 2) 1care first and foremost about your underlying intelligenceand that's what I value you for (a fixed mindset message).In short, we worried that praise for intelligence would putstudents in a fixed mindset with all of its vulnerabilities.When we thought about the children with a growthmindset, we thought about their focus on effort and strategies,and how this focus on process allowed them to remainmotivated and effective in the face of setbacks. So we wonderedwhether praise for effort or strategies would promotea growth mindset with its motivation and resilience.We then conducted research to test these ideas. Westudied fifth graders and kindergarteners. We studied childrenin inner city schools, suburban schools, and ruralschools. And we found the same thing in each case.After students received intelligence praise, they adopteda fixed mindset. They rejected a challenging task they couldlearn from, instead selecting the task that would makethem look smart. When they hit difficulty and made errors,they lost confidence in their ability - now they thoughtthey were not smart - and ended up performing poorly.Students who were praised for their effort entered agrowth mindset. They wanted the challenge, they maintainedtheir confidence and enjoyment in the face of difficulty,and they ended up performing far better, even whenthe task was an IQtest.MAKE NO MISTAKE - CHILDREN LOVED THE INTELLIGENCE PRAISE.THEY SMILED BROADLY AND SEEMED PROUD OF THEMSELVES. IT REALLYMADE THE TESTERS FEEL AS THOUGH THEY HAD GIVEN THE CHILDRENSOMETHING VALUABLE.BUT OUR FINDINGS TOLD A DIFFERENT STORY.There was one more intriguing finding. Students whowere praised for their intelligence later lied about theirscores. This means that errors were so humiliating thatthey could not own up to them.Make no mistake - children loved the intelligence praise.They smiled broadly and seemed proud of themselves. Itreally made the testers feel as though they had given thechildren something valuable. But our findings told a differentstory. It was praising the student's process - whichcould be their effort, strategies, concentration, choices,persistence - that helped them remain motivated, confident,and effective. (For ideas about ways to deliver processpraise, see my book Mindset.)Can a growth mindset be taught directly?GROWTH MINDSET PROGRAMSBOOST ACHIEVEMENTThree recent studies (by Joshua Aronson, by CatherineGood, and by my group) have shown that teaching studentsa growth mindset results in increased motivation,better grades, and higher achievement test scores. Over aseries of sessions, students were taught that their brainsCANADIAN EDUCATION ASSOCIATION I EDUCATION CANADA 9

form new connections every time they learn, and that overtime they can become smarter. Students were very excitedby the idea that they could influence their brains. Theywere also shown how to apply this idea to their schoolwork.Whether the students were in junior high school orat an elite university, those who received this message outperformedstudents in the control groups (even when thestudents in the control groups received excellent training instudy skills). They also reported a greater investment inlearning, and teachers reported noticeable changes in thesestudents' desire to work hard and learn.These benefits were especially important for studentswho are subject to negative stereotypes - girls in math orAfrican-American students. The effects of stereotypes werereduced when students believed that their abilities couldbe developed. For example, in one study the gender gap inmath was greatly reduced when girls were taught thegrowth mindset.Measuring Students' M1ndsets (for students 10 and older)Read each sentence below and then circle the one number that shows how much you agreewith it. There are no right or wrong answers.* You have a certain amount of intelligence, and you really can't do much to change It.*Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6CONCLUSIONMany teachers see evidence for a fixed mindset every year.The students who start out at the top of their class end upat the top, and the students who start out at the bottomend up there. Research by Falko Rheinberg shows thatwhen teachers believe in fixed intelligence, this is exactlywhat happens. It is a self-fulfilling prophecy. However,when teachers hold a growth mindset, many students whostart out lower in the class blossom during the year andjoin the higher achievers.As educators, we want all of the students we teach toprofit from our efforts. A growth mindset - ours and theirs- helps students to seek learning, to love learning, and tolearn effectively. ICAROL S. DWECK is the Lewis and Virginia Eaton Professorof Psychology at Stanford University and the author ofMindset: The New Psychology of Success, recently published byRandom House. She is a leading researcher in the area ofstudent motivation, having received numerous grants andawards for her work. Her research has also been featured inTime magazine, New York magazine, the New York Times, theBoston Globe, the Washington Post, the Manchester Guardian,and on 20/20, the Today show, and the BBC.A list of readings related to this article is available on theCEA website at www.cea-ace.caW Your intelligence is something about you that you can't change very much.*Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6E You can learn new things, but you can't really change your basic intelligence.*Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6H No matter who you are, you can change your intelligence a lot.Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6* You can always greatly change how intelligent you are.Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6II No matter how much intelligence you have, you can always change it quite a bit.Strongly Agree 1 Agree 2 Mostly Agree 3Mostly Disagree 4 Disagree 5 Strongly Disagree 6*These three fixed-mindset Items can be used alone.Do you know someone who hasmade an outstanding educationalcontribution to sustainable pratices?Find out how you can nominate them for theExcellence in Education Award for Promotion ofSustainable Practices.Canada Mortgage and Housing Corporation (CMHC)has established the Excellence in Education Award forPromotion of Sustainable Practices to honoureducators engaged in innovative environmentaleducation and community development.EligibilityEducators in all secondary and post-secondaryinstitutions across Ontario are eligibleNomination deadline is May 31, 2007 (Nominationforms available at www.cmhc.ca keyword: Excellencein Education)Nominate someone today!Go to www.cmhc.ca keyword: Excellence in Educationfor award criteria and nomination form.Canad70^C.%10 EDUCATION CANADA I CANADIAN EDUCATION ASSOCIATION

IOF A GOOD THING?We encouragethem, wecelebrate them,and we givethem stickersand stars. Butare we doingmore harmthan good?By SamanthaCleaversk Korean eighth graders,"Are you good at math?"and chances are they'llsay they aren't. Ask anAmerican, and you'lllikely get an enthusiasticresponse. In a recentstudy, only six percentof Korean eighth graders consideredthemselves excellent math students,compared with 39 percent of Americaneighth graders. Yet the Korean studentsscored far better in math than theirAmerican peers.We've taught our children since birthto believe they can do anything theychoose, from starring in the school playto mastering long division. All that selfconfidence,however, hasn't producedmore capable students. The BrookingsInstitution 2006 Brown Center Reporton Education finds that countries inwhich families and schools emphasizeself-esteem for students-America forexample-lag behind the cultures thatdon't focus on how students feel aboutthemselves.For decades our culture has concentratedon teaching self-esteem first,learning second. In the late 1980s, aCalifornia government task force foundno connection between low self-esteemand societal ills, such as drug use, teenpregnancy, and school underachievement.Still, California forged ahead witha self-esteem education plan. Today,raising children's self-esteem continuesto be a primary goal in the classroom,and a goal of parents at home.Downplaying grades, praisingchildren for minimal effort, or usingneutral-colored green or purple pensto comment on written work seemsharmless enough, but we may be takingaway the sense of satisfaction andpride that comes from genuine achievement.Jason Walsh, a special educationteacher in Washington, D.C., witnessedINSTRUCTOR SEPTEMBER/OCTOBER 2007 31

cover storythis firsthand during his school's fifthgradegraduation ceremonies. Some studentsreceived as many as 14 differentawards. "The majority of the studentsdidn't know what their awards reallymeant," says Walsh. The honors "didn'treinforce a specific achievement-but asense of entitlement and of being great."The long-term impact of this rah-rahmentality is already apparent. In 2004,according to Jean Twenge, author ofGeneration Me, 70 percent of Americancollege freshmen reported their academicability as "above average." But,once ego-inflated students get to college,they're more likely to drop out, saysTwenge, when their skewed sense of selfand overconfidence affects their abilityto make decisions.A growing contingent of expertsagree that in the classroom, self-esteemshould be an outcome, not a method."Self-esteem," says Robert Brooks,Harvard Medical School faculty psychologist,"is based on real accomplishments."It's all about letting kids shinein a realistic way.FEEL-GOOD ACADEMICSThere is a correlation between selfesteemand grades; studies have shownthat high grades do lead to high selfesteem.But rather then praising childrenfor every effort along the way,we should encourage them to focus onachieving particular goals and applaudthat achievement. The danger of toomuch praise is that children may turntheir focus to how good they feel andhow to get more praise, rather thanon what they're learning. "All childrenshould be held accountable," says KarenBernstein, music teacher at Howard B.Mattlin Middle School in Long Island,New York. "We shouldn't worry aboutsetting boundaries for kids becauseof their feelings." In her classroom,Bernstein makes children responsiblefor their actions, so that when they doachieve, "they feel good."If a student's confidence isn't built onhis or her actual abilities, failure can bedevastating. Walsh worries that his students'sense of greatness may lead to a"psychological crash and burn" becauseThe Myths ofSelf-EsteemIt's easy to fall for these poputar truisms.Can you separate fact from fiction?they don't understand why they failed.Walsh's hunch is on track, according toexperts; children who work solely forpraise don't feel intrinsic satisfaction.Even if a child feels competent, saysJohn Shindler, associate professor ofeducation at California State University,it's not real competence if it's rooted inconstant praise.Furthermore, students absorbed withtheir own sense of self often have troublecompleting difficult tasks. One studyshows that adolescent girls have lowerself-esteem than adolescent boys. Whenconfronted with a low score, the girlsare more likely to work to improve theirperformance, while the boys are morelikely to give up and change activities."The real issue," says Rick Weissbourd,who teaches at the Harvard GraduateSchool of Education, "is getting kids todevelop a sense of self-efficacy, alongwith real competencies and skills. Selfesteemwill follow."WHAT TEACHERS CAN DOThe latest findings on self-esteem canbe dispiriting for teachers. After all,who wouldn't want to make studentsfeel good about themselves? Rather thantear down your "Shoot for the Stars"poster, reassess your priorities. Makesure that your primary focus is on studentperformance and improvement,rather than how kids feel about themselves.Emphasize effort and specificcharacter traits, such as persistence,helpfulness, and consideration.Students need to see that achievementis related to the effort they put out.Establish clear expectations and rubricsthat students can use to achieve the outcome,with effort as one of the criteriafor success. If a student has trouble withlarge goals, says Bernstein, break thetask into achievable chunks.It's all about realism, adds Twenge.Instead of focusing on how great thestudents are, highlight students' realstrengths. Teach them that none of uscan be good at everything all the time.Weissbourd recommends that for eachchild in your class, identify three of hisor her strengths and then make aplan to highlight those strengths.32 INSTRUCTOR SEPTEMBER/OCTOBER 2007

cover storyAsk your students, what can theylearn from this situation? What canthey contribute? What do they want toachieve this week, month, or year? Thenhelp them create goals and ways toachieve those goals.Focus on actions and real charactertraits, not "special" and "great." Thebest kind of praise, says Weissbourd,"communicates a specific achievementand the importance of effort in thatachievement." When you correct children,focus on what they can do betternext time and show them how.Don't shelter children from failure.Children who are shielded from failureare not prepared to deal with adversity.When students fail, tell them exactlywhy and provide clear ways for themto succeed the next time. To break theice, suggests Brooks, tell your studentsabout a time when you failed in schooland how you recovered.A CLASSROOM FULL OF VIPSOf course, you are working with childrenwho have already been raised in aself-esteem world. They may think thatthey deserve recognition regardlessof how they perform and believe theyshould be considered first. Kids who actout don't have poor self-esteem, saysTwenge. Instead, they often think thatthey're the most important person inthe room and that everyone else is gettingin their way.When a child is disruptive, you needto figure out why the child is actingthat way and work with him or herto fix it. William Ricks, a teacher inSussex, Virginia, asks his students towalk with him in the hall to talk withhim about their behavior. "I try to findthe root of their attitude," says Reid,"and then I talk to them about humility."To order, call 1-800-724-6527;fax 1-800-560-65; or visitwww.scholasflc.com/pofessionalAddressing students' needs is crucialfor behavior and academics. Once students'"social and emotional needs aremet," says Bernstein, "they will be morelikely to work harder."Focusing on praise and avoiding criticismmakes everybody feel good. Butchildren who have high self-esteem maybecome rude and uncooperative whenthey're criticized. Still, "don't try toCIRCLE NO.35 ON TEACHER ACTION CARD34 INSTRUCTOR SEPTEMBER/OCTOBER 2007

protect students from failure," saysJennifer Crocker, Claude M. SteeleCollegiate Professor of Psychology at theUniversity of Michigan. Instead, whenwe make failure a learning experienceand not a threat, students' self-esteemisn't on the line, and they're more opento taking constructive criticism.Young children are naturally narcissistic,and teaching them self-esteemkeeps them focused on themselves,instead of thinking about others."Narcissism separates you from otherpeople," says Twenge, while true "selfesteembrings you into connectionwith other people." In the long term,narcissism has been linked with aggressionand poor relationships, while connectingchildren to other people has apositive affect on behavior. And, saysCrocker, children learn more whenthey're supportive of others.More important than self-esteem,says Weissbourd, is a child's maturity,or the ability to be aware of others,coordinate other people's needs withhis or her own, and regulate intensefeelings. By rewarding our students'social successes, such as helping theirpeers, being good community members,and listening, we increase their genuineself-esteem and improve their behavior.Allowing children to help around theclassroom, says Brooks, increases their"realistic self-esteem [because childrenare] making a positive difference in thelife of someone else."Walsh has worked with studentswho have inflated egos and no senseof responsibility or respect. Too muchself-esteem, he says, "creates a sense ofentitlement. I'm not saying that childrendon't need reinforcement, but you haveto make sure that you develop a realistic,practical, and consistent behaviorplan." When we focus on building students'self-control, sense of belonging,and competence, we create moreself-esteem than we (to if we dole outconstant praise. "Genuine self-esteem,"says Shindler, is "a set of unconsciousself-beliefs, formed over a lifetime,reflecting our perceptions and abilities,our ability to love, and how we attributecausality for the events in our lives." L1Achievement through Literacy & Character DevelopmentIncrease reading achievement and inspire your students to want to learn withVoices Reading-the first comprehensive literacy program to systematicallyaddress the academic, social, and emotional needs of your children.Co-authored by renowned literacy expert, Dr. Catherine Snow,Voices Reading has been proven to increase student achievement inreading and writing while at the same time increasing positive behaviorand student engagement in the classroom.k-S77-1 T6tCýTo request a Voices Reading Program Overview, contactZaner-Bloser Customer Service at 800.42L3018.7r Zaner-BloserLThe Language A3 s and Reading CompanyCIRCLE NO. 43 ON TEACHER ACTION CARD

Even GeniusesLet’s give students learning tasks that tell them,“You can be as smart as you want to be.”Carol S. DweckWe can all agree that meaningful schoolwork promotesstudents’ learning of academic content.But why stop there? I believe that meaningfulwork can also teach students to love challenges,to enjoy effort, to be resilient, and tovalue their own improvement. In other words, we can design andpresent learning tasks in a way that helps students develop a growthmindset, which leads to not just short-term achievement but alsolong-term success.Why Foster a Growth Mindset?During the past several decades, my colleagues and I have conductedresearch identifying two distinct ways in which individualsview intelligence and learning. Individuals with a fixed mindsetbelieve that their intelligence is simply an inborn trait—they havea certain amount, and that’s that. In contrast, individuals with agrowth mindset believe that they candevelop their intelligence overtime (Blackwell, Trzesniewski,& Dweck, 2007; Dweck,1999, 2007).These two mindsetslead to different schoolbehaviors. For one thing,when students view intelligenceas fixed, they tendto value looking smartabove all else. They maysacrifice important opportunitiesto learn—even those thatare important to their futureacademic success—if thoseopportunities require them to riskperforming poorly or admittingdeficiencies. Students with a growthmindset, on the other hand, viewchallenging work as an opportunityto learn and grow. I have seen students© STEFANIE FELIX

Work Hardwith a growth mindset meet difficultproblems, ones they could not solve yet,with great relish. Instead of thinkingthey were failing (as the students with afixed mindset did), they said things like“I love a challenge,” “Mistakes are ourfriends,” and “I was hoping this would beinformative!”Students with a fixed mindset donot like effort. They believe that if youhave ability, everything should comenaturally. They tell us that when theyhave to work hard, they feel dumb.Students with a growth mindset, incontrast, value effort; they realize thateven geniuses have to work hard todevelop their abilities and make theircontributions.Finally, students with a fixed mindsettend not to handle setbacks well.Because they believe that setbacks calltheir intelligence into question, theybecome discouraged or defensive whenthey don’t succeed right away. Theymay quickly withdraw their effort,blame others, lie about their scores,or consider cheating. Students witha growth mindset are more likely torespond to initial obstacles by remaininginvolved, trying new strategies, andusing all the resources at their disposalfor learning.Creating a Culture of Risk TakingTeachers who strive to design challenging,meaningful learning tasksmay find that their students responddifferently depending on the students’© THE GRANGER COLLECTIONASCD / WWW. ASCD. ORG 17

The highest reward for a person’s toil is not whatMeaningful learning tasksgive students a clear sense ofprogress leading to mastery.© SUSIE FITZHUGHassumptions about intelligence. Studentswith a growth mindset may tacklesuch work with excitement, whereasstudents with a fixed mindset may feelthreatened by learning tasks that requirethem to stretch or take risks.To prepare students to benefit frommeaningful work, therefore, teachersneed to create a growth-mindset culturein the classroom. One way to createsuch a culture is by providing the rightkinds of praise and encouragement. Myresearch has shown that praising studentsfor the process they have engagedin—the effort they applied, the strategiesthey used, the choices they made,the persistence they displayed, and soon—yields more long-term benefitsthan telling them they are “smart” whenthey succeed.Teachers should also emphasize thatfast learning is not always the deepestand best learning and that studentswho take longer sometimes understandthings at a deeper level. Students canlearn about many historical figures whowere not regarded as “fast” learners inchildhood. Albert Einstein swore thathe was slow to learn and that’s whyhe pondered the same questions yearafter year—with, as we know, excellentresults.Some teachers teach their studentsabout the different mindsets directly.(To learn about a growth mindset curriculumthat my colleagues and I havecreated, go to www.brainology.us.)Teachers may illustrate the conceptof the growth mindset by having theirstudents write about, and share withone another, something they used to bepoor at and are now very good at.In one class, for example, the studentswere astounded to learn that theschool’s baseball star used to be ineptat baseball and only became proficientafter much practice. Such discussionsencourage students not to be ashamedto struggle with something before theyare good at it.Teachers can also ask their studentsto choose an area in which they wouldlike to improve and then to establish apersonal goal that would be a big reachfor them. For example, a student who istypically afraid of criticism might decideto seek critical feedback on her next artproject; an algebra student strugglingto understand absolute values mightcommit to watching a YouTube videoon how to solve linear absolute valueequations, and then teach the processto his classmates; a student who lacks18 EDUCATIONAL LEADERSHIP / SEPTEMBER 2010

they get for it, but what they become by it. —John Ruskinphysical confidence might join a sportsteam; or a shy student might approachother students she would like tobefriend. Students can share their plansand even help one another enhancetheir skills and reach their goal.Another strategy is to have studentswrite a letter to a struggling studentexplaining the growth mindset, tellingthe struggler not to label himself orherself, and giving the student advice onimprovement strategies to try.Through such exercises, teachersare transmitting crucial information—telling students that they view themall as having intelligence that they canchoose to develop. The teachers are alsocommunicating that their role is not tojudge who is smart and who is not, butto collaborate with students to makeeveryone smarter.Building a Growth MindsetWithin a classroom culture that supportsa growth mindset, teachers can designmeaningful learning tasks and presentthem in a way that fosters students’resilience and long-term achievement.Emphasize Challenge, Not “Success”Meaningful learning tasks need to challengeevery student in some way. It iscrucial that no student be able to coastto success time after time; this experiencecan create the fixed-mindsetbelief that you are smart only if you cansucceed without effort.To prevent this, teachers can identifystudents who have easily mastered thematerial and design in-class assignmentsthat include some problems or exercisesthat require these students to stretch.This way, the teacher will be close athand to guide students if necessary andget them used to (and ultimately excitedabout) the challenging work. Someteachers have told me that after a while,students begin to select or create challengingtasks for themselves.When presenting learning tasks tostudents, the teacher should portraychallenges as fun and exciting, whileportraying easy tasks as boring and lessuseful for the brain. When studentsinitially struggle or make mistakes, theteacher should view this as an opportunityto teach students how to try differentstrategies if the first ones don’twork—how to step back and thinkabout what to try next, like a detectivesolving a mystery.Suppose that a student has attempteda math problem but is now stuck. Theteacher can say, “OK, let’s solve thismystery!” and ask the student to showthe strategies he or she has tried so far.As the student explains a strategy, theteacher can say, “That’s an interestingstrategy. Let’s think about why it didn’twork and whether it gives us someclues for a new path. What should wetry next?”When, perhaps with the teacher’sguidance, the student finds a fruitfulstrategy, the teacher can say “Great!You tried different ways, you followedthe clues, and you found a strategythat worked. You’re just like SherlockHolmes, the great detective. Are youready to try another one?” In this way,the teacher can simultaneously gaininsight into what the student doesand does not understand and teachthe student to struggle through knottyproblems.Give a Sense of ProgressMeaningful learning tasks give studentsa clear sense of progress leading tomastery. This means that students cansee themselves doing tasks they couldn’tdo before and understanding conceptsthey couldn’t understand before.Work that gives students a sense ofimprovement as a result of effort givesteachers an opportunity to praise studentsfor their process. That is, teacherscan point out that the students’ effortswere what led to the progress andimprovement over time.Some teachers make students’progress explicit by giving pre-tests atthe beginning of a unit that purposelycover material students do not know.When students compare their inevitablypoor performance on these pre-testswith their improved performance onunit post-tests, they get used to the ideathat, with application, they can becomesmarter.Homework is an especially importantcomponent of an instructional program© SUSIE FITZHUGHA S C D / w w w . a s c d . o r g 19

that enhances students’ sense of learningand progress. Homework assignmentsshould not feel like mindless, repetitiveexercises; rather, they should presentnovel problems for students to solve,require them to apply what they’velearned in new ways, or ask them tostretch to the next level.For example, suppose that studentsare learning about the rise andfall of civilizations. Their homeworkassignment might be to apply theirlearning by designing a civilization thatwould either thrive (by building inpositive factors) or implode (by buildingin risk factors). They can write the storyof their civilization and what happenedto it. Or suppose students were studyingShakespeare’s sonnets. For homework,they could write a sonnet to the personor animal of their choice in the style ofShakespeare.Grade for GrowthThe way teachers evaluate their students’work can also help studentsdevelop a growth mindset. At one highschool in Chicago, when students don’tmaster a particular unit of study, theydon’t receive a failing grade—instead,Fast learning is notalways the deepestand best learning.they get a grade of Not Yet. Students arenot ashamed of that grade because theyknow that they’re expected to masterthe material, if not the first time, thenthe next time, or the next.The word “yet” is valuable and shouldbe used frequently in every classroom.Whenever students say they can’t dosomething or are not good at something,the teacher should add, “yet.”Whenever students say they don’t like acertain subject, the teacher should say,“yet.” This simple habit conveys the ideathat ability and motivation are fluid.Some teachers my colleagues and Iwork with tell us that they’ve shiftedtheir grading system to consider moregrowth-mindset criteria, so that nostudent can coast to an A and studentswho struggle and improve getcredit for their effort. One school basesone-fourth of each student’s grade ongrowth-mindset factors, thus rewarding© SUSIE FITZHUGHstudents who challenge themselves, areresilient in the face of difficulty, andshow clear improvement over time.Other schools give a separate grade forchallenge-seeking, effort, and resilience.Of course, for that grade to be effective(and not just a consolation prize),teachers need to have reinforced thevalue of these qualities daily throughoutthe school year.What if a student puts in great effortbut does not improve? The teacherneeds to factor in the effort but thenwork with the student to figure outwhat the impasse was and how thestudent can break through that impasse.Long-Term SuccessMeaningful work not only promoteslearning in the immediate situation, butalso promotes a love of learning andresilience in the face of obstacles. Thiskind of meaningful work takes place inclassrooms in which teachers praise thelearning process rather than the students’ability, convey the joy of tacklingchallenging learning tasks, and highlightprogress and effort. Students who arenurtured in such classrooms will havethe values and tools that breed lifelongsuccess. ELReferencesBlackwell, L., Trzesniewski, K., & Dweck,C. S. (2007). Implicit theories of intelligencepredict achievement across anadolescent transition: A longitudinalstudy and an intervention. Child Development,78(1), 246–263.Dweck, C. S. (1999). Self-theories: Their rolein motivation, personality, and development.Philadelphia: Psychology Press.Dweck, C. S. (2007). The perils andpromises of praise. Educational Leadership,65(2), 34–39.Carol S. Dweck is the Lewis and VirginiaEaton Professor of Psychologyat Stanford University and the authorof Mindset: The New Psychologyof Success (Random House, 2006);dweck@stanford.edu.20 E d u c a t i o n a l L e a d e r s h i p / S e p t e m b e r 2 0 1 0

Copyright of Educational Leadership is the property of Association for Supervision & Curriculum Developmentand its content may not be copied or emailed to multiple sites or posted to a listserv without the copyrightholder's express written permission. However, users may print, download, or email articles for individual use.

SHOULD SCHOOLS TRY TOBOOST SELF ESTEEM?Beware the dark sideBY ROY F. BAUMEISTERRoy F. Baumeister is the Elsie Smith professor of psychology at Case Western ReserveUniversity in Cleveland, Ohio. For a fuller discussion of the relationship between self-esteemand violence, see "Relation of Threatened Egotism to Violence and Aggression: The Dark Sideof High Self-Esteem," by Roy F. Baumeister, Laura Smart, and Joseph M. Boden(Psychological Review, 199(5, Vol. 103, No. 1).“WE MUST raise children's self-esteem!" How often has this sentimentbeen expressed in recent years in schools, homes, and meeting rooms aroundthe United States? The sentiment reflects the widespread, well-intentioned,earnest, and yet rather pathetic hope that if we can only persuade our kids tolove themselves more, they will stop dropping out, getting pregnant, carryingweapons, taking drugs, and getting into trouble, and instead will startachieving great things in school and out.Unfortunately, the large mass of knowledge that research psychologistshave built up around self-esteem does not justify that hope. At best, high selfesteemis a mixed blessing whose total effects are likely to be small andminor. At worst, the pursuit of high self-esteem is a foolish, wasteful, andself-destructive enterprise that may end up doing more harm than good.Writers on controversial topics should acknowledge their biases, and so letme confess mine: I have a strong bias in favor of self-esteem. I have beenexcited about self-esteem ever since my student days at Princeton, when I firstheard that it was a topic of study. Over the past two decades I have probablypublished more studies on self-esteem than anybody else in the United States(or elsewhere). It would be great for my career if self-esteem could doeverything its boosters hope: I'd be dining frequently at the White House andadvising policymakers on how to fix the country's problems.It is therefore with considerable personal disappointment that I mustreport that the enthusiastic claims of the self-esteem movement mostly rangefrom fantasy to hogwash. The effects of self-esteem are small, limited, andnot all good. Yes, a few people here and there end up worse off because theirself-esteem was too low. Then again, other people end up worse off becausetheir self-esteem was too high. And most of the time self-esteem makessurprisingly little difference.Self-esteem is, literally, how favorably a person regards himself or herself.It is perception (and evaluation), not reality. For example, I think the worldwould be a better place if we could all manage to be a little nicer to eachother. But that's hard: We'd all have to discipline ourselves to change. Theself-esteem approach, in contrast, is to skip over the hard work of changingour actions and instead just let us all think we're nicer. That won't make theworld any better. People with high self-esteem are not in fact any nicer thanpeople with low self-esteem—in fact, the opposite is closer to the truth.High self-esteem means thinking well of oneself, regardless of whetherthat perception is based on substantive achievement or mere wishful thinkingand self-deception. High self-esteem can mean confident and secure—but itcan also mean conceited, arrogant, narcissistic, and egotistical.

A recent, widely publicized study dramatized the fact that self-esteemconsists of perception and is not necessarily based on reality. In aninternational scholastic competition, American students achieved the lowestaverage scores among all participating nationalities. But the American kidsrated themselves and their performance the highest. This is precisely whatcomes of focusing on self-esteem: poor performance accompanied by plentyof empty self-congratulation. Put another way, we get high self-esteem asinflated perceptions covering over a rather dismal reality.Looking ahead, it is alarming to think what will happen when thisgeneration of schoolchildren grows up into adults who may continue thinkingthey are smarter than the rest of the world—while actually being dumber.America will be a land of conceited fools.All of this might fairly be discounted if America were really sufferingfrom an epidemic of low self-esteem, such as if most Americanschoolchildren generally had such negative views of themselves that theywere unable to tackle their homework. But that's not the case. On the contrary,as I'll explain shortly, self-esteem is already inflated throughout the UnitedStates. The average American already regards himself or herself as aboveaverage. At this point, any further boosting of self-esteem is likely toapproach the level of grandiose, egotistical delusions.Benefits of Self-EsteemLet us begin with the positive consequences of high self-esteem. Muchhas been claimed, but very little has been proven. Some years ago Californiaformed a task force to promote self-esteem, and its manifesto was filled withoptimistic assertions about how raising self-esteem would help solve most ofthe personal and social problems in the state. Here is a sample of its rhetoric:"the lack of self-esteem is central to most personal and social ills plaguing ourstate and nation," and indeed self-esteem was touted as a social vaccine thatmight inoculate people "against the lures of crime, violence, substance abuse,teen pregnancy, child abuse, chronic welfare dependency, and educationalfailure."'Such rhetoric is especially remarkable in light of another fact. That sametask force commissioned a group of researchers to assemble the relevant factsand findings about self-esteem. Here is what the experts in charge of theproject concluded from all the information they gathered: "The news mostconsistently reported, however, is that the associations between self-esteemand its expected consequences are mixed, insignificant, or absent." In short,self-esteem doesn't have much impact.Even when the occasional study does link low self-esteem to someproblem pattern, there is often a serious chicken-and-egg ambiguity aboutwhich comes first. For example, if someone showed that drug-addictedpregnant unmarried school-dropout teenagers with criminal records have lowself-esteem, this might mean only that people stop bragging after they messup their lives. It would not prove that low self-esteem caused the problems.The few researchers who have tried to establish causality have usuallyconcluded that self-esteem is mainly an outcome, not a cause. At best there isa mutual influence of spiraling effects.To be sure, there are some benefits of high self-esteem. It helps peoplebounce back after failure and try again. It helps them recover from trauma andmisfortune. In general, high self-esteem makes people feel good. Low self-

esteem accompanies various emotional vulnerabilities including depressionand anxiety. (Again, though, there is no proof that low self-esteem causesthese problems, or that raising self-esteem will prevent them.)Children who do well in school have slightly higher self-esteem than thosewho do poorly. Unfortunately the effect is small, and in fact anyone whobelieves in the value of education should wish for a stronger effect simply onthe basis that successful students deserve higher self-esteem. Across multiplestudies, the average correlation between grades and self-esteem is .24, whichmeans about 6 percent of the variance. In other words, moving from the veryhighest self-esteem scores to the very lowest would yield about a 6 percentdifference in school performance. A small increase in self-esteem, such asmight be produced by a school program aimed at boosting self-esteem, wouldprobably make only a 1 percent difference or less. And even that assumes thatself-esteem is the cause, not the effect, contrary to many indications. To theextent that it is school success or failure that alters self-esteem and not theother way around, any independent effort to raise self-esteem would have noeffect at all on school performance.Once again I must say how disappointing I've found these facts to be.Self-esteem is not altogether useless but its benefits are isolated and minor,except for the fact that it feels good. When I embarked on a career in researchon self-esteem, I had hoped for a great deal more.The Dark Side of High Self-EsteemThe very idea that high self-esteem could have bad consequences strikessome people as startling. The self-esteem movement wants to present selfesteemas having many good and no bad effects. But very few psychologicaltraits are one-sidedly good, and those few are mostly abilities (likeintelligence or self-control). High self-esteem can certainly cause its share ofproblems. If you pause to recall that the category of high self-esteem includespeople who think they are great without necessarily being great, thisconclusion may seem less startling.A large, important study recently adopted a novel approach to separatingself-esteem from all its cause and correlates. The researchers measured howeach individual rated himself or herself compared to how that person wasrated by others who knew him or her. They were particularly interested in thecategory of people with inflated self-esteem—the ones who rated themselveshigher than their friends rated them. This after all, is where the self-esteemmovement leads: Concentrate on getting kids to think well of themselves,regardless of actual accomplishments. The researchers had no difficultyfinding plenty of students who fit that category. They are, in a sense, the starproducts and poster children of the self-esteem movement.And what were they like? The researchers' conclusions did not paint anencouraging picture of health, adjustment, or success. On the contrary, thelong-term outcomes of these people's lives found above average rates ofinterpersonal and psychological problems. A second study, with laboratoryobservations of live interactions, showed these people to be rather obnoxious.They were more likely than others to interrupt when someone else wasspeaking. They were more prone to disrupt the conversation with angry andhostile remarks. They tended to talk at people instead of talking to or withthem. In general, they irritated the other people present. Does any of thissound familiar? This is what comes of inflated self-esteem.

The picture is one of a self-centered, conceited person who is quick toassert his or her own wants but lacks genuine regard for others. That may notbe what the self-esteem movement has in mind, but it is what it is likely toproduce. In practice, high self-esteem usually amounts to a person thinkingthat he or she is better than other people. If you think you're better than others,why should you listen to them, be considerate, or keep still when you want todo or say something?Over the past several years, I have been writing a book on evil andviolence (Evil: Inside Human Violence and Cruelty, to be published byFreeman this fall). Given my longstanding interest in self-esteem, I naturallywanted to acknowledge any part that it plays. Various pundits and so-calledexperts have long asserted that low self-esteem causes violence, but I've hadenough experience with self-esteem to know that I'd better check the datarather than relying on vague generalizations and ostensibly "common"knowledge.Two graduate students and I reviewed literally hundreds of studies on thetopic. What we found was so surprising that in addition to my book, werecently published a lengthy article in psychology's most eminent journal, thePsychological Review. We combined evidence from all spheres of violencewe could find: murder, assault, rape, terrorism, bullies, youth gangs,repressive governments, tyranny, family violence, warfare, oppression,genocide, and more.We concluded that the idea that low self-esteem causes violence is simplyand thoroughly wrong. It is contradicted by a huge mass of information andevidence. People with low self-esteem are generally shy, humble, modest,self-effacing individuals. Violent perpetrators—from Hitler, Hussein, andAmin, down to the common wife-beater or playground bully—are decidedlynot like that.If anything, high self-esteem is closer to the violent personality. Mostperpetrators of violence are acting out of some sense of personal superiority,especially one that has been threatened or questioned in some way. I am notsaying that high self-esteem, per se, directly causes violence. Not all peoplewith high self-esteem become violent. But violent people are a subset ofpeople with high self-esteem. The main recipe for violence is threatenedegotism—that is, a belief in personal superiority that is challenged,questioned, or "dissed" by somebody else. Inflated self-esteem often leads tothat pattern.Consider some of the evidence. In the first place, whenever there are twogroups with different levels of self-esteem, the more egotistical group isnearly always the more violent one. The most familiar example is gender.Men have higher self-esteem and higher rates of violence. When self-esteemfluctuates, the risk of violence rises with the favorable views of self, such asin manic-depressive illness. Indeed, people who are intoxicated with alcoholshow increases in self-esteem and increases in violent tendencies.A recent study found that nowadays many homicides occur in connectionwith other crimes such as robbery, but in the remaining cases the homicide isoften the result of an altercation that begins with challenges and insults, inwhich someone's favorable self opinion is disputed by the other person. Theperson who feels he (or less often she) is losing face in the argument mayresort to violence and murder.Even within samples of offenders, it appears that indicators of egotism candiscriminate violent and troublesome tendencies, and it is the favorable view

of self that are linked to the worse actions. A group of researchersadministered the California Psychological inventory to young men (in theirlate teens) on parole. The researchers were able to predict future paroleviolations (recidivism) better than previous attempts. Among the traits thatpredicted high recidivism were being egotistical and outspoken (as well as"touchy" which suggests being easily offended). Meanwhile being modest andunassuming (associated with low self-esteem) were among the traits linked tobe least likely to violate parole. These results all seem fit the view linkingfavorable views of self to violent tendencies.Aggression starts in childhood, and bullies are the most notable examples.They are of particular importance because childhood bullies have been foundto be four times more likely than other children to engage in serious criminalbehavior during their subsequent adult life. Dan Olweus is an expert who hasstudied bullies for years, and he recently summarized the conclusions that hisprogram of research has yielded. Unlike victims of bullying (who showmultiple indications of low self-esteem), the bullies themselves seemedrelatively secure and free from anxiety. In contrast to a fairly commonassumption among psychologists and psychiatrists, we have found noindicators that the aggressive bullies (boys) are anxious and insecure under atough surface," said Olweus, adding that multiple samples and methods hadconfirmed the conclusion, and concluding that bullies "do not suffer frompoor self-esteem."'One of the most earnest and empathic efforts to understand the subjectiveexperience of committing crimes was that of sociologist Jack Katz. Homicideas well as assault emerged in his study as typically caused by threats to theoffender's public image. In Katz’s view, the offender privately holds apositive view of self, but the eventual victim impugns that view and implicitlyhumiliates the offender, often in front of audience. The response is unplannedviolence resulting in injury or death. Katz insisted that feelings of beinghumiliated are quickly transformed into rage. He argued that many men feelthat almost anyone can judge them and impugn their esteem, whereas forwomen’s self-esteem is most heavily invested in their intimate relationships—with the result that men will attack strangers while women mainly just murdertheir intimate partners, because only the partners can threat their self-esteemto a sufficient degree to provoke such a violent response.Another example of the relationship between inflated self-esteem andviolence focuses on juvenile delinquency. The classic study by Glueck andGlueck compared juvenile delinquents against a matched sample ofnondelinquent boys. Although the study was an early one and has beencriticized on methodological grounds, it benefited from a large sample andextensive work, and nearly all of their findings have been replicated bysubsequent studies. The Glueck and Glueck study did not measure self-esteemdirectly (indeed it antedated most modern self-esteem scales), but there wereplenty of related variables. The pattern of findings offers little to support thehypothesis that low self-esteem causes delinquency. Delinquent boys weremore likely than controls to be characterized as self-assertive, sociallyassertive, defiant, and narcissistic, none of which seems compatible with lowself-esteem. Meanwhile, the delinquents were less likely than the comparisongroup to be marked by the factors that do indicate low self-esteem, includingsevere insecurity, feelings of helplessness, feelings of being unloved, generalanxiety (a frequent correlate of low self-esteem), submissiveness, and fear of

failure. Thus, the thoughts and actions of juvenile delinquents suggested thatthey held quite favorable opinions of themselves.It is useful to look for convergences between the Gluecks' study and morerecent studies of youthful violence, not only because of the seminal nature ofthe Gluecks' work, but also because their data were collected several decadesago and on an almost entirely white sample, unlike more recent studies.Converging findings thus confer especially high confidence in conclusionsthat can be supported across time and ethnicity.One of the most thorough research projects on youth gangs was that ofMartin Sanchez Jankowski, whose work involved 10 years, several cities, and37 gangs. Although as a sociologist he was disinclined to use self-esteem orpersonality factors as explanatory constructs, his study did furnish severalimportant observations. Jankowski specifically rejected the notion that actingtough is a result of low self-esteem or feelings of inadequacy. In his words,"There have been some studies of gangs that suggest that many gang membershave tough exteriors but are insecure on the inside. This is a mistakenobservation" (p. 27). He said that for many members, the appeal of the gang isthe positive respect it enjoys in the community as well as the respectfultreatment from other gang members, which he found to be an important normin nearly all gangs he studied. He said most gang members "expressed astrong sense of self-competence and a drive to compete with others.” Whenthey failed, they always blamed something external rather than personalinadequacy or error. This last observation is especially relevant becauseseveral controlled studies have shown that it is characteristic of high selfesteemand contrary to the typical responses of people with low self-esteem.Recently I appeared on a radio talk show. The hostess seemed to havedifficulty accepting the conclusion that low self-esteem is not a cause ofviolence, possibly because she had swallowed the propaganda line that allgood things come from high self-esteem. To explain our findings, I offeredthe example of the Ku Klux Klan. The KKK has long advocated beliefs inwhite superiority and has turned violent in response to effort to extend fullequality to black citizens (thereby eroding the superior status of whites). Ithought KKK violence was a good, clear example of threatened egotism.For a moment the hostess seemed to see the point but then she jumpedback on the self-esteem band wagon. "What about deep down inside?" sheasked. I inquired whether she thought that Klansmen believed that they, aswhites were inferior to blacks, which would fit the low self-esteem view. Shebalked at the word "inferior" but offered that the violent Klansmen believedeep down inside that they are "not superior"—in other words, equal—toblacks.I didn't know what to say to this basically loony argument. Her theory thatKlan violence could be traced to a 'deep down’ inner belief that blacks areequal to whites has two parts, both of which are bizarre: first that members ofthe KKK truly believe in racial equality, and second, that belief in racialequality causes violence. It struck me that attempts to defend the self-esteemmovement against the facts end up having to make such preposterousassertions.Although this particular hostess's idea was absurd she was invoking apoint that the proponents of self esteem have on occasion raised as a possiblyvalid defense. When obnoxious or socially undesirable acts are performed byegotistical people, thus contradicting the belief that high self-esteem isgenerally good, some propose that these obnoxious individuals must secretly

have low self-esteem. Indeed, the editorial reviewers who evaluated ourarticle on violence for the Psychological Review insisted that we tackle thistheoretical question head-on in the final published version of the paper.There are two main reasons to reject the "hidden low self-esteem" view.The first is that plenty of researchers have tried and failed to find anyindications of this allegedly hidden low self-esteem. It's not for lack of trying,and indeed it would be quite a feather in any researcher's cap to show thatactions are caused by low self-esteem hidden under a veneer of high selfesteem.Studies of childhood bullies, teen gang members, adult criminals, andvarious obnoxious narcissists keep coming to the same conclusion: "We'veheard the theory that these people have low self-esteem or a negative selfimageunderneath, but we sure can't find any sign of it."The other reason is even more compelling. Suppose it were true (which itdoes not seem to be) that some violent people have high self-esteem on thesurface but low self-esteem inside. Which view of self (the surface veneer orthe hidden one) would be the one responsible for violence? We already knowthat genuine low self-esteem, when not hidden, does not cause violence.Hence one would have to say that low self-esteem is only linked to violencewhen it is hidden. That means that the crucial cause of violence is what ishiding the secret insecurity—which means that the "veneer" of high selfesteemis the cause, and so we are back anyway to the position that egotism isthe cause.There isn't space here to exhaust the dark side of high self-esteem, but letme touch on a few other features. People with high self-esteem are lesswilling than others to heed advice, for obvious reasons—they usually thinkthey know better. (Whether children with inflated self-esteem are less willingto listen to teachers is one possible implication of this, but to my knowledgethis has not yet been studied.) They respond to failure by blaming everyoneand everything but themselves, such as a flawed test, a biased or unfairteacher, or an incompetent partner. They sometimes extend their favorableself-opinion to encompass people close to or similar to themselves, butunfortunately this often translates into prejudice and condescension towardpeople who differ from them. (High self-esteem is in fact linked to prejudiceagainst out-groups.) Finally, when their egotism is threatened, they tend toreact irrationally in ways that have been shown to be risky, self-defeating, andeven self-destructive.Boosting Self-Esteem: The Problem of InflationMost (though not all) of the problems linked to high self-esteem involveinflated self-esteem, in the sense of overestimating oneself. Based on theresearch findings produced in laboratories all over North America, I have noobjection to people forming a sober, accurate recognition of their actualtalents and accomplishments. The violence, the self-defeating behaviors, andthe other problems tend to be most acute under conditions of threatenedegotism, and inflated self-esteem increases that risk. After all, if you really aresmart, your experiences will tend to confirm that fact, and so there's not muchdanger in high self-esteem that is based on accurate recognition of yourintelligence. On the other hand, if you overestimate your abilities, reality willbe constantly showing you up and bursting your bubble, and so your (inflated)self-opinion will be bumping up against threats—and those encounters lead todestructive responses.

Unfortunately, a school system that seeks to boost self-esteem in generalis likely to produce the more dangerous (inflated) form of self-esteem. Itwould be fine, for example, to give a hard test and then announce the top fewscores for general applause. Such a system recognizes the successful ones,and it shows the rest what the important criteria are (and how much they mayneed to improve). What is dangerous and worrisome is any procedure thatwould allow the other students to think that they are just as accomplished asthe top scorers even though they did not perform as well. Unfortunately, theself-esteem movement often works in precisely this wrong-headed fashion.Some students will inevitably be smarter, work harder, learn more, andperform better than others There is no harm (and in fact probably somepositive value) in helping these individuals recognize their superioraccomplishments and talents. Such self-esteem is linked to reality and henceless prone to causing dangers and problems.On the other hand, there is considerable danger and harm in falselyboosting the self-esteem of the other students. It is fine to encourage them towork harder and try to gain an accurate appraisal of their strengths andweaknesses, and it is also fine to recognize their talents and accomplishmentsin other (including nonacademic) spheres, but don't give them positivefeedback that they have not earned. (Also, don't downplay the importance ofacademic achievement as the central goal of school, such as by suggestingthat success at sports or crafts is just as good.) To encourage the lowerperformingstudents to regard their performance just as favorably as the toplearners—a strategy all too popular with the self-esteem movement—is atragic mistake. If successful, it results only in inflated self-esteem, which isthe recipe for a host of problems and destructive patterns.The logical implications of this argument show exactly when self-esteemshould be boosted. When people seriously underestimate their abilities andaccomplishments, they need boosting. For example, a student who falselybelieves she can't succeed at math may end up short-changing herself andfailing to fulfill her potential unless she can be helped to realize that yes, shedoes have the ability to master math.In contrast, self-esteem should not be boosted when it is already in theaccurate range (or higher). A student who correctly believes that math is nothis strong point should not be given exaggerated notions of what he canaccomplish. Otherwise, the eventual result will be failure and heartbreak.Along the way he's likely to be angry, troublesome, and prone to blameeverybody else when something goes wrong.In my years as an educator I have seen both patterns. But which is morecommon? Whether boosting self-esteem in general will be helpful or harmfuldepends on the answer. And the answer is overwhelmingly clear. Far, farmore Americans of all ages have accurate or inflated views of themselvesthan underestimate themselves. They don't need boosting.Dozens of studies have documented how inflated self-esteem is. Researchinterest was sparked some years ago by a survey in which 90 percent of adultsrated themselves "above average" in driving ability. After all, only half canreally be above average. Similar patterns are found with almost all goodqualities. A survey about leadership ability found that only 2 percent of highschool students rated themselves as below average. Meanwhile, a whopping25 percent claimed to be in the top 1 percent! Similarly, when asked aboutability to get along with others, no students at all said they were belowaverage.

Responses to scales designed to measure self-esteem show the samepattern. There are always plenty of scores at the high end and plenty in themiddle, but only a few straggle down toward the low end. This seems to betrue no matter which of the many self-esteem scales is used. Moreover, thefew individuals who do show the truly low self-esteem scores probably sufferfrom multiple problems that need professional therapy. Self-esteem boostingfrom schools would not cure them.Obviously there's precious little evidence of low self-esteem in suchnumbers. By definition, plenty of people are in reality below average, butmost of them refuse to acknowledge it. Meanwhile large numbers of peopleclearly overestimate themselves. The top 1 percent can really only contain 1percent, not the 25 percent who claim to belong there. Meanwhile, theproblem that would justify programs aimed at boosting self-esteem—peoplewho significantly underestimate themselves—is extremely rare.ConclusionWhat is to be done? In response to the question about whether schoolsshould boost self-esteem, my answer is: Don't bother. Efforts at boosting selfesteemprobably feel good both for students and for teachers, but the realbenefits and positive consequences are likely to be minor. Meanwhile,inflated self-esteem carries an assortment of risks and dangers, and so effortsto boost self-esteem may do as much harm as good, or possibly even more.The time, effort; and resources that schools put into self-esteem will not bejustified by any palpable improvements in school performance, citizenship, orother outcomes.There is one psychological trait that schools could help instill and that islikely to pay off much better than self-esteem. That trait is self-control(including self-discipline). Unlike self-esteem, self-control (or lack thereof) isdirectly and causally involved in a large set of social and personal problems."Addiction, crime, violence, unwanted pregnancy, venereal disease, poorschool performance, and many other problems have self-control failure as acore cause. Also unlike self-esteem, self-control brings benefits to both theindividual and society. People with better self-control are more successful(socially and academically), happier, and better adjusted, than others. Theyalso make better parents, spouses, colleagues, and employees. In other words,their self-control benefits the people close to them.Indeed, I am convinced that weak self-control is a crucial link betweenfamily breakdown and many social problems. Study after study has shownthat children of single parents show up worse than average on almost everymeasure, ranging from math achievement tests to criminal convictions. Mostsingle parents I know are loving, dedicated, hard-working individuals, but alltheir energy goes toward providing food and shelter and their children's otherbasic needs. It seems to take a second parent to provide the supervision andconsistent rule enforcement that foster self-control in the child.How much the schools can do to build self-control is unclear. Still, justrecognizing the priority and value of self-control will help. Obviously, selfcontrolis not something that is instilled directly (as in a "self-control class")but rather should be cultivated like a cluster of good habits in connection withregular academic work, especially in the context of clear, consistentenforcement of academic and behavioral standards. The disciplinary andacademic culture of a school should be aimed at recognizing and encouraging

the self-control of individual students, including rewarding good self-controland punishing its failures or absences. With each new plan, policy, orprocedure, school officials might pause to ask "Will this help strengthen selfcontrol?"instead of "Might this hurt anybody's self-esteem?"In the long run, self-control will do far more for the individuals and forsociety as a whole than will self-esteem. Moreover, self-control gives peoplethe ability to change and improve themselves, and so it can bring aboutchanges in substantive reality, not just in perception. And if one can makeoneself into a better person, self-esteem is likely to increase too. Raising selfcontrolmay thus end up boosting self-esteem—but not in the dangerous orsuperficial ways that flourish now.My final message to all the people working in today's schools and seekingto help the next generation get a good start is, therefore, as follows: Forgetabout self-esteem, and concentrate on self-control._________________Note: The format of this document has been altered to make for easier readability, to remove extraneouscharacters, and to provide more space for annotations.The original document can be accessed at this link.

Social Psychology of Education (2007) 10:303–330 © Springer 2007DOI: 10.1007/s11218-007-9020-4Comparing self-esteem and perceived controlas predictors of first-year college students’academic achievementROBERT H. STUPNISKY 1,∗ , ROBERT D. RENAUD 1 , RAYMONDP. PERRY 1 , JOELLE C. RUTHIG 2 , TARA L. HAYNES 1and RODNEY A. CLIFTON 11 University of Manitoba, Winnipeg, Canada2 University of North Dakota, Fargo, USA(Received 30 June 2006; Accepted in final from 31 January 2007)Abstract. Due to its widespread popularity, self-esteem is continually being promoted tostudents despite limited empirical support for its effectiveness in improving their academicachievement. As a result, constructs that are potentially more salient to academic performance,such as perceived control, have gone relatively unnoticed. Although past researchhas examined the link between students’ academic achievement and either their self-esteemor perceived control, few studies have compared both constructs simultaneously to elucidatewhich one is more important to academic success. This longitudinal study directlycontrasted the effects of self-esteem and perceived control on the academic performance of802 first-year college students. After accounting for incoming ability (high school grades),age, and gender, a structural equation model showed perceived control positively predictedstudents’ GPA. In contrast, the predictive effect of self-esteem on GPA was non-existent.Findings indicate that compared to self-esteem, perceived control is a more powerful predictorof first-year college students’ GPA. Implications for utilizing educational interventionsto boost perceived control among college students are discussed.Key words: self-esteem; perceived control; college students; academic achievement;structural equation modeling1. IntroductionFor some time now, educational institutions have sought to bolster students’self-esteem to increase their academic success based on the commonbelief that self-esteem is universally beneficial and advantageous inall areas of life. Despite self-esteem’s strong link to increased life satisfactionand happiness (see Baumeister, Campbell, Kruegger, & Vohs, 2003for a review; Diener & Diener, 1995), empirical evidence indicates thatself-esteem is a weak, unreliable predictor of academic achievement (Bachman& O’Malley, 1986; Crocker & Luhtanen, 2003; Maruyama, Rubin, &∗ Author for correspondence: e-mail: umstupnl@cc.umanitoba.ca

304 ROBERT H. STUPNISKY ET AL.Kingsbury, 1981; Rosenberg, Schooler, & Schoenbach, 1989; Skaalvik &Hagtvet, 1990). In addition to empricial shortcomings, certain trends donot support the assumed benefits of self-esteem. For example, althoughthe overall level of college students’ self-esteem has increased over thepast several decades (Twenge & Campbell, 2001), drop-out rates of firstyearcollege students remain high at approximately 27% (Geraghty, 1996).Because indicators of academic aptitude, specifically high school gradesand SAT scores, have been found to account for only about 20% offirst-year college students’ GPAs (Zwick & Sklar, 2005), individual differenceconstructs are believed to be major contributors to college students’achievement. Although college students would presumably desire both highself-esteem and academic success, the question remains whether bolsteringself-esteem is the most effective means of fostering their academicsuccess.Given the lack of evidence linking self-esteem to achievement, researchershave recently focused on linking an alternative individual differenceconstruct to college students’ academic performance, namely perceivedcontrol. First-year college students experience a challenging transition fromhigh school to college that involves increased emphasis on performance,heightened competition, more frequent failure, unfamiliar academic tasks,new social networks, and critical career choices. These novel and oftendaunting developments can make some students feel “out of control” leadingthem to perceive college as a low-control environment (Perry, 1991;Perry, Hall, & Ruthig, 2005).Furthermore, if a student were then to experience a poor performance,such as failing a test, they may be more likely to attribute thatperformance to factors such as an overly demanding professor, or anexceedingly difficult course. Uncontrollable attributions such as these leadto decreased feelings of responsibility, negative affect, decreased motivation,and poor performance in the future (Weiner, 1985, 1995). Asa result of these challenging first-year circumstances and potential maladaptiveattributional patterns, perceived control is receiving increasedattention as an important predictor of college students’ academic achievement.Despite consistent empirical support for perceived control’s influenceon academic achievement (Perry, Hladkyj, Pekrun, Clifton, & Chipperfield,2005; Perry, Hladkyj, Pekrun, & Pelletier, 2001; Ruthig, Perry et al.,2007), it appears that educators and administrators will continue to focuson self-esteem to explain students’ academic achievement until researcherselucidate a better alternative. The current study compared the effects ofcollege students’ self-esteem and perceived control on their academic performancein order to clarify which construct best predicts their academicsuccess.

SELF-ESTEEM AND PERCEIVED CONTROL 3051.1. SELF-ESTEEM AND ACADEMIC ACHIEVEMENTSelf-esteem is commonly regarded as the positive or negative attitude a personhas towards the concept of the self (Rosenberg, 1965), and is viewedby some psychologists as a fundamental human motive that we strive toprotect and enhance (Rosenberg et al., 1989). Baumeister et al. (2003)remarked that North Americans view self-esteem as a highly desirable qualityand a central psychological source of positive behavior. This view ofself-esteem is endorsed by organizations such as the National Associationfor Self-esteem whose mission statement is to “promote awarenessof and provide vision, leadership and advocacy for improving the humancondition through the enhancement of self-esteem” (http://www.self-esteemnase.org/boardmembers.shtml,Mission section). The breadth of self-esteemis also evident in the scientific literature as it has become one of the mostresearched constructs in all of psychology (Judge, Erez, Bono, & Thoresen,2002).Regrettably, this “one-size-fits-all” perspective has led to the overuseof self-esteem as a means of bolstering academic success in the absenceof strong empirical support. For example, it is commonly believed thatstudents with greater self-esteem attain higher grades. Baumeister et al.(2003) theorized that this belief stems from the perception that people withhigher, relative to lower, self-esteem have greater aspirations, are more persistentin the face of failure, and are less likely to succumb to feelings ofincompetence and self-doubt. Based on this perception, numerous educationalprograms and interventions have been implemented in an attemptto increase students’ grades by boosting their self-esteem. These programshave been included in massive interventions such as Head Start for preschoolers,Follow Through for primary grade students, and Upward Boundfor high school students (Scheirer & Kraut, 1979). In recent years, a largenumber of programs designed to enhance self-esteem have also proliferated(Haney & Durlak, 1998), particularly reflected in numerous collegecounseling homepages touting the benefits of having high self-esteem andsuggesting strategies for increasing it. Unfortunately, these programs wereimplemented without strong or consistent empirical evidence linking selfesteemwith academic performance.Much of the research examining the relationship between self-esteemand academic achievement has been conducted among adolescents andhigh school students. These studies show a small but consistent positivecorrelation between self-esteem and academic achievement ranging fromr = .10 to .30 (see Baumeister et al., 2003 for a review; Pottebaum, Keith,& Ehly, 1986). However, correlation must not be confused with causationand the results from studies that tested a causal connection betweenthese two variables have been mixed. The causal effect predicted by selfesteemadvocates would be for students with high self-esteem to have

306 ROBERT H. STUPNISKY ET AL.higher grades than students with low self-esteem. Studies testing this directcausal effect via structural equation modeling have generally shown selfesteemto be either a weak or a non-significant predictor of academic performance(Bachman & O’Malley, 1986; Maruyama et al., 1981; Rosenberget al., 1989; Skaalvik & Hagtvet, 1990).Although evidence for a direct effect is inconsistent, the testing of otherlinkages between self-esteem and academic achievement have been moresuccessful. Specifically, several studies tested the reverse causal directionand found that academic achievement has a significant positive effect onself-esteem (Rosenberg et al. 1989, Ross & Broh, 2000). Also, a study byLui, Kaplan, and Risser (1992) found an indirect effect of self-esteem onachievement whereby low self-esteem resulted in increased deviance, psychologicaldistress, decreased motivation, illness, and absence from school,in turn, resulting in poor academic performance for students.Comparatively few studies have examined the link between self-esteemand academic achievement among college students, yet the findings areequally inconsistent to the results involving adolescents and high schoolstudents. Woodward and Suddick (1992), for example, found a positive correlationbetween self-esteem and college GPA, however Demo and Parker(1988) found a non-significant correlation. Crocker and Luhtanen (2003)used regressions to examine self-esteem as a predictor of college students’academic achievement and found that the level of self-esteem did notsignificantly predict GPA. However, they did find that self-esteem negativelypredicted students’ self-reported academic problems, such as lowergrades than desired and dissatisfaction with school. To summarize, thesmall positive correlations between self-esteem and academic achievementfor adolescents, high school students, and college students have not beenclarified empirically in studies examining the direct causal effect of selfesteemon academic achievement. Instead, the strongest indication frompast research is that high self-esteem is likely to be the result, rather thanthe cause, of academic success.1.2. PERCEIVED CONTROL AND ACADEMIC ACHIEVEMENTRotter (1966) was one of the first researchers to study perceived controlby differentiating between individuals who explained events according tointernal versus external causes (i.e., locus of control). More recently, perceivedcontrol has been defined as a person’s belief in his or her capacityto influence and predict some aspect of the environment (Perry, 1991; Perryet al., 2005; see Skinner, 1996 for a review). The prevailing view in theresearch literature is that individuals with higher, relative to lower, levels ofperceived control are advantaged in many ways, such as experiencing lowerlevels of stress and depression (Garber & Seligman, 1980), improved health(Thompson, 1981), and longer life span (Chipperfield, 1993). Furthermore,

SELF-ESTEEM AND PERCEIVED CONTROL 307several meta-analyses on students ranging from Grade 1 to college haveconsistently found internal locus of control beliefs, which are analogousto high perceptions of control, to be positively associated with higheracademic achievement (Findley & Cooper, 1983; Kalechstein & Nowicki,1997).Perceived control is relevant to college students’ academic achievementfor several reasons. Perry (1991, 2003) has argued that students’ responsibilityfor their achievement is increasingly emphasized as they progressthrough the education system, making perceived control especially importantamong college students. For example, a number of tasks carried outfor high school students by their teachers and principles are no longermonitored in college, such as schedule planning, checking class attendance,inquiries about late or missing assignments, and seeking out struggling studentsto offer one-on-one instruction. College students are also requiredto complete a number of new and challenging tasks including registration,choosing a major, selecting appropriate courses, self-guided studying, challengingcourse assignments, applying for scholarships and awards, securingfinancial resources, and finding appropriate housing. These tasks arefrequently carried out under circumstances that can make college studentsfeel out of control, such as increased emphasis on success–failure, heightenedacademic competition, increased pressure to excel, more frequent failure,unfamiliar tasks, new social networks, and critical career choices. As aconsequence, many students perceive college as a low-control environment.Those students who see themselves as “in control” of these tasks often takeinitiatives to succeed, while those who feel unable to control these circumstancestake less responsibility, tend to be more vulnerable to failure, andmay subsequently drop out of college (Perry et al., 2005).Weiner’s (1985; 1995) Attribution Theory of Motivation and Emotionalso provides insight into the importance of perceived control to collegestudents’ academic achievement. Weiner posits that students’ explanations,or attributions, for their failures are pivotal to both their achievementstriving and their academic performance. When a student with low perceivedcontrol, for example, performs below his or her expectations (e.g.,unexpectedly failing a test), the student is likely to attribute the failure touncontrollable factors, such as “I’m not smart enough to succeed in college”or “the professor made the test too difficult”. According to Weiner,uncontrollable attributions such as these are linked to low perceptions ofresponsibility, feelings of shame, decreased motivation, ultimately resultingin diminished academic performance. Conversely, a student with a highersense of control is more likely to attribute an unexpected failure to controllablecauses, such as “I didn’t try hard enough” or “I didn’t use effectivestudy strategies”. Controllable attributions typically contribute to increasedperceptions of responsibility, feelings of guilt, motivation to study, ultimatelyfostering successful performance. Thus, having a high perception of

308 ROBERT H. STUPNISKY ET AL.control is extremely important to first-year college students because it helpsthem maintain an appropriate level of motivation to persevere in difficultsituations.College students’ perceptions of control have been studied in relationto several indicators of motivation and achievement-striving as revealed invarious measures of academic achievement. Perry, Hladkyj, Pekrun, andPelletier (2001), for example, found that students with higher, relative tolower, perceptions of control obtained better course grades, reported havinghigher motivation, exerting more effort, experiencing less boredom andanxiety, using self-monitoring strategies more often, and felt more controlover their lives in general. A three-year follow up of the same students(Perry, Hladkyj et al., 2005) found that those higher in perceived controlobtained better GPAs three years later and withdrew from fewer coursesthan students lower in control. A study by Ruthig, Perry et al. (2007) providedadditional support for the benefits of perceived control on academicachievement, as students high in perceived control obtained significantlybetter course grades and GPAs than students low in perceived control.In addition to achievement outcomes, perceived control has been linkedto other factors that contribute to academic success in college. Cassidyand Eachus (2000) found that students with higher perceptions of controlused more effective study strategies than students with lower perceptionsof control, resulting in greater academic achievement. Schönwetter, Perry,and Struthers (1993) found that high-control students felt more pride andless shame concerning their course performances than low-control students.Furthermore, several studies note the beneficial effects of perceived controlon academic achievement as moderated by individual student characteristics.For example, among students with high perceptions of controlover academic outcomes, those who were preoccupied with failure obtainedhigher grades (Perry et al., 2001) and dropped fewer courses over a 3-year period (Perry, Hladkyj et al., 2005), than all other students. Together,these findings indicate that perceived control consistently influences academicachievement both directly and indirectly through students’ cognitions,emotions, and motivation.1.3. COMPARING THE EFFECTS OF SELF-ESTEEM AND PERCEIVEDCONTROL ON ACADEMIC ACHIEVEMENTRoss and Broh (2000) suggested that an overlap between self-esteem andperceived control is intuitive because people who perceive themselves asbeing in control of their lives also tend to see themselves as people ofworth. Alternatively, they acknowledged that individuals may view themselvesas good people who are well-respected by others, yet believe thatmost outcomes are beyond their control. This latter alternative suggests

SELF-ESTEEM AND PERCEIVED CONTROL 309that the impacts of perceived control and self-esteem within a particularcontext (e.g., academics) may differ.Many empricial studies have examined the association between selfesteemand perceived control (or locus of control). In a meta-analysis,Judge et al. (2002) identified 47 studies containing a correlation betweenself-esteem and locus of control and found that the average correlation wasr = .52 (95% CI = .44–.59). However, only a few studies compared theimpact of self-esteem and perceived control on academic outcomes. A fewexceptions include Ross and Broh’s (2000) longitudinal study that examinedperceived control and self-esteem among a sample of Grade 10 studentsand found that only perceived control positively predicted students’course grades in Grade 12. Similarly, Cappella and Weinstein (2001) examinedstudents with a low Grade 8 reading proficiency and found studentswho reported a high internal locus of control were more likely to improvetheir reading proficiency by Grade 12 than students who had an externallocus of control; alternatively, self-esteem was found to have no significanteffect on the students’ reading proficiency. Finally, a study by Mooney,Sherman, and Lo Presto (1991) found that, for first-year college students’,academic locus of control and self-esteem were each positively correlatedwith college adjustment; however, the relationship between locus of controland college adjustment was much stronger. It should be noted that noneof these studies used a scale specifically designed to measure perceived control,a limitation that the current study will overcome. Nonetheless, previousresearch suggests that perceived control is a more powerful predictor ofvarious academic outcomes than self-esteem, however, self-esteem and perceivedcontrol have yet to be directly compared as predictors of college students’academic achievement (e.g., GPA).In brief, a review of the self-esteem and perceived control literaturerevealed three substantial issues to be addressed in the current study:(1) the relationship between self-esteem and academic achievement isinconsistent, yet programs designed to improve college students’ academicperformance have tended to focus on enhancing self-esteem; (2) the relationshipbetween perceived control and academic achievement is strongerand more consistent than that of self-esteem, yet perceived control has notbeen embraced to the same degree as self-esteem in attempting to improvecollege students’ academic performance; and (3) although students’ levelsof perceived control and self-esteem appear to be related, few studies havecompared their influence on the academic achievement of college studentsto test which construct is the most relevant to their academic success.To address these issues, our study compared self-esteem and perceivedcontrol as predictors of first-year college students’ academic performance.The key methodological strengths of this study include: (1) longitudinaldata collected over three points in time spanning from high school to theend of students’ first year of college; (2) employing structural equation

310 ROBERT H. STUPNISKY ET AL.modeling to test causal paths while statistically controlling for students’incoming ability (high school grades), gender, and age; (3) the use of establishedmulti-item scales for measuring self-esteem and perceived controlto create latent variables; and (4) utilizing objective measures of academicachievement (high school grades and college GPAs) obtained from institutionalrecords.2. Method2.1. PARTICIPANTSParticipants were 802 first-year students (498 females, 304 males; age 17–22)from a large, mid-western research-1 university who were enrolled in atwo-semester introductory psychology course. The students were recruitedat the start of the academic year and volunteered for the study to obtaincredits in their introductory psychology course. 12.2. MEASURES2.2.1. High School GradesStudents’ final average percentage in their core college entrance courses(Math, English, Chemistry, Physics) from their last year of high schoolwere obtained from institutional records at the end of the academic year(M = 77.63%, SD = 8.01, range = 56.00–99.00). High school performancewas included in the analyses to control for potential differences in students’aptitude when they entered college (Perry et al., 2001; Perry, Hladkyi et al.,2005; Ruthig, Haynes et al., 2007), which may affect self-esteem, perceivedcontrol, and/or GPA.2.2.2. DemographicsAge (M = 18.23, SD = 0.57) was included in the analysis because it positivelycorrelates with self-esteem: older students have higher self-esteemthan younger students (Twenge & Campbell, 2001; Woodard & Suddick,1992). Age has also been found to correlate with academic achievement, asolder students have more academic success (Clifton, Perry, Adams-Stubbs,& Roberts, 2004). Gender was included because female college studentstend to report lower self-esteem than male students (Demo & Parker,1988).2.2.3. Self-esteemStudents’ self-esteem was measured using the Rosenberg (1965) Self-esteemScale (RSES). The RSES is a well-established measure with high reliabilityand construct validity (Robins, Hendin, & Trzesniewki, 2001). Thescale contains five positively-worded items, such as “I feel that I have a

SELF-ESTEEM AND PERCEIVED CONTROL 311number of good qualities”; and five negatively-worded items, for example“At times I think that I am no good at all”, measured on a five-point Likertscale (1 = Strongly disagree; 5=Strongly agree). The negatively-wordeditems were reverse coded so that for each of the 10 items a higher scoreindicated higher self-esteem (see Appendix A for complete scale).2.2.4. Perceived ControlStudents’ perceived control was assessed using Perry et al.’s (2001) PerceivedAcademic Control Scale, which contains eight items measured on afive-point Likert scale (1 = Strongly disagree; 5=Strongly agree). The scaleincludes four positively-worded items, such as “The more effort I put intomy courses, the better I do in them”, and four negatively-worded items, forexample “My grades are basically determined by things beyond my controland there is little I can do to change that”. The negatively-worded itemswere reverse coded so that a high score indicated a high level of perceivedcontrol for each item.2.2.5. Grade Point AverageAcademic performance in college was measured using students’ cumulativegrade point averages (GPA) obtained from institutional records after theconclusion of the academic year. 2 GPA constitutes students’ average gradeattained in all courses completed during their first year of college and isexpressed as a numeric value based on the following institutional cut-offvalues: 0 = F, 1 = D, 2 = C, 2.5 = C+, 3 = B, 3.5 = B+, 4 = A, 4.5 = A+ (M =2.66, SD = .91, range = .14–4.43).2.3. PROCEDUREApproximately one month into the academic year (October), the data werecollected from students in groups of 10–35 in a classroom setting. Eachstudent confirmed a willingness to participate in the study by signing bothan informed consent and grade release form. This was followed by thestudents completing a questionnaire measuring their demographics, selfesteem,and perceived control. Students were allowed as much time as theyneeded to complete the questionnaire, typically requiring 30–40 min. Afterthe conclusion of the academic year the students’ high school grades andGPAs were obtained from institutional records.3. Results3.1. RATIONALE FOR ANALYSESOur analyses were conducted in three steps. First, confirmatory factoranalyses (CFA) tested the validity of the measurement models (i.e., the

312 ROBERT H. STUPNISKY ET AL.relations of the observed/measured variables to the latent/unmeasured variablesof self-esteem and perceived control). This was done before assessingthe relationship between the study variables using correlations and thefull structural model to reduce potential measurement problems, as recommendedby Marsh, Byne, and Yeung (1999). Second, correlations werecalculated between high school grades, age, gender, the latent variables forself-esteem and perceived control, and GPA. Correlations were computedin order to examine the zero-order relationships between the study variablesprior to a comparative analysis of their impact on academic achievement.Third, a structural equation model (SEM) was used to contrast students’self-esteem and perceived control as predictors of their academic achievement,taking into account high school grades, age, and gender. This analysiswas chosen because SEM provides estimates of measurement error in themodel, and thus, provides estimates of the true relations among the variablesof self-esteem, perceived control, and GPA. All analyses were conductedusing AMOS (version 5.01; Arbuckle, 2003) using the maximumlikelihood estimation method.3.1.1. Step 1: The Measurement ModelsTwo criteria were used to test the quality of the model’s fit in additionto the traditional chi-square (χ 2 ) goodness of fit measure. The comparativefit index (CFI; Bentler, 1990) compares the improved fit of the hypothesizedmodel to a null model that assumes the observed variables in themodel are uncorrelated (i.e., the “independence model”). CFI values rangefrom 0 (poor fit) to 1.0 (perfect fit) with values that approximate or exceed.95 representing a well-fitting model (Hu & Bentler, 1999). The root meansquare error of approximation (RMSEA; Browne & Cudeck, 1993) representsthe fit between the implied and the population covariance matrix. Inaddition, the RMSEA is a parsimony-adjusted index in that it becomesworse as the hypothesized model becomes more complex. Unlike the CFI,the RMSEA could be considered a “badness-of-fit” measure with lowervalues (i.e., closer to 0) being more desirable (Kline, 2005). Generally, aRMSEA ≤.05 represents a good fit (Browne & Cudeck, 1993).Self-esteem. To account for a potential method effect resulting fromincluding positively and negatively worded items, we estimated the covariancesbetween error terms for each of the positively and negatively wordeditems, a procedure which is consistent with previous CFAs on the RSES(Ross & Broh, 2000; Tomas & Oliver, 1999; Whiteside-Mansell & Corwyn,2003). As the increased number of estimated parameters yielded a justidentifiedmodel, one of the residual covariances was set to zero (specificallybetween the two positively worded items SE1 and SE7, as was doneby Whiteside-Mansell & Corwyn, 2003) to allow for over-identification.The fit results support the inclusion of the self-esteem construct in subsequentanalyses, χ 2 (16) = 58.03, p

SELF-ESTEEM AND PERCEIVED CONTROL 313Table I. Zero-order correlations for variables in study1 2 3 4 5 61. Gender a –2. Age .05 –3. High school grades −.12 ∗ −.12 ∗ –4. Self-esteem .17 ∗ −.01 .08 –5. Perceived control .01 −.04 .24 ∗ .43 ∗ –6. GPA −.07 .00 .70 ∗ .12 ∗ .31 ∗ –a Gender coded female = 1, male = 2.∗ Correlation is significant at p

314 ROBERT H. STUPNISKY ET AL.Figure 1. Structural model with standardized coefficients. Note. All coefficients greaterthan or equal to .09 were significant at the p

SELF-ESTEEM AND PERCEIVED CONTROL 315Table II. Baseline model standardized parameter estimatesfor self-esteem and perceived controlSelf-esteem Perceived control Error varianceSE1 .56 .50SE2 .61 .40SE3 .76 .34SE4 .48 .57SE5 .72 .52SE6 .68 .53SE7 .65 .58SE8 .70 .80SE9 .63 .88SE10 .70 .79PC1 .45 .70PC2 .40 .50PC3 .70 .57PC4 .48 .34PC5 .57 .47PC6 .67 .32PC8 .66 .36Note. See Appendix A for item wordings.self-esteem, predict GPA. Age positively predicted GPA (β = .09, p

316 ROBERT H. STUPNISKY ET AL.Table III. Standardized regression coefficientsand R 2 for study variables predicting GPAβGender .00Age.09 ∗High school grades.68 ∗Self-esteem .02Perceived control.13 ∗Adjusted R 2 .52∗ p

SELF-ESTEEM AND PERCEIVED CONTROL 317was correlational, it is equally likely that perceived control was affectingself-esteem, or that both were affected by an unknown third variable. Forthis reason, self-esteem’s indirect effect on GPA should be interpreted withcaution.The second issue was that although the relationship between perceivedcontrol and academic achievement is more substantive and consistent thanthat of self-esteem and achievement, perceived control has not been consideredto the same extent as self-esteem in explaining academic performance.In keeping with previous studies (Perry et al., 2001; Perry, Hladkyjet al., 2005; Ruthig, Haynes et al., 2007), we found that students’ level ofperceived control had a significant positive effect on their GPAs, an effectthat has been overshadowed by the perceived impact of the more popularconstruct of self-esteem. In addition to the main results, our analysesrevealed several other intriguing findings. For example, students’ highschool grades were the strongest predictor of students’ GPA, which is consistentwith previous research on the transition of high school students tocollege (Perry et al., 2001; Perry, Hladkyj et al., 2005; Ruthig, Haynes etal., 2007). Impressively, perceived control was found to significantly predictstudents’ GPA even after accounting for this measure of students’ incomingability, along with age and gender. Another noteworthy result was that theeffect of perceived control on GPA was much larger than the effects of ageand gender on GPA (β = .09 and β =−.01, respectively), which are bothhighly intuitive predictors of academic performance. These results furtherunderscore the robustness of perceived control’s effect on college students’academic achievement.The third issue was implicit in the first two; although students’ levelsof perceived control and self-esteem tend to be related, few studies havecompared their unique effects on academic achievement to determine whichconstruct is the most relevant to students’ academic performance. Our resultsindicated that perceived control had a larger impact on college students’ academicachievement than did self-esteem. Although the lack of a direct effectof self-esteem on academic achievement (GPA) may be counter to “commonwisdom”, this finding is consistent with previous empirical research involvinghigh school students (Ross & Broh, 2000). Overall, our results show that ifthe goal of higher educational institutions is to successfully guide studentsthrough academic programs, of self-esteem is being overvalued as its effectson academic achievement are questionable. Alternatively, perceived control isbeing undervalued given its beneficial effects on college students are strongerand more consistent.4.1. SELF-ESTEEM IN COLLEGEThere are several plausible explanations for why self-esteem was not asignificant predictor of college students’ GPA in our study. First, according

318 ROBERT H. STUPNISKY ET AL.to Rosenberg (1965), a person who has high self-esteem considers him orher self a good person who is respected by others. Based on this definition,self-esteem has clear benefits for psychological well-being, such as greaterhappiness and life satisfaction (Baumeister et al., 2003; Diener & Diener,1995); however it is not clear why self-esteem would necessarily impact students’academic success in college. A student who has low self-esteem couldbe as motivated as a student who has high self-esteem to study hard, useeffective test taking strategies, and write high quality papers in order toattain a higher GPA. Alternatively, the finding that previous achievement(high school grades) increases self-esteem is a more intuitive effect, as goodgrades are in fact positive evaluations by others (i.e., professors) which hasbeen found to be a means of boosting self-esteem (Rosenberg et al., 1989;Ross & Broh, 2000). This implies that high self-esteem is more likely to bethe result than the cause of academic success, and suggests increasing students’self-esteem would be ineffective in improving their academic performance.A second possible explanation for self-esteem’s non-significant impacton GPA was revealed by our literature review. We found the majorityof research on students’ self-esteem has focused on elementary and highschool students, whereas the majority of research on perceived control hasinvolved college students. This suggests that researchers recognize that theage or educational level of students may impact which individual differenceconstruct influences their academic achievement. In fact, some researchershave suggested that the effect of self-esteem on academic achievement ismore salient among younger students and declines later in high school andsubsequent years (Bachman & O’Malley, 1977; Demo & Parker, 1988).This trend may occur because of differences in how students perceivetheir educational environments. As mentioned previously, the first yearof college can be viewed as a low-control environment (Perry, 1991,2003) and consequently perceived control may be more influential, whichwas supported by our results. In elementary and high school, however,the environment is made more controllable by teachers and principalswho follow up on late homework assignments, provide more one-on-oneteaching for weaker students, allow for make-up tests, etc. Consequently, inpre-college educational environments students may be more likely to perceivethemselves as in control. What may hinder younger students fromperforming well academically is an underdeveloped sense of self-esteem,which has been found to result in low happiness and life satisfaction(Baumeister et al., 2003; Diener & Diener, 1995), that distracts them fromtheir school work. On the other hand, older students are more likely tohave developed a relatively mature sense of self-esteem (Twenge & Campbell,2001; Woodard & Suddick, 1992), that has less of an effect on theiracademic performances in college. Although the study by Ross and Broh(2000) on high school students also found perceived control to have a

SELF-ESTEEM AND PERCEIVED CONTROL 319stronger influence on achievement than self-esteem, if a similar model wastested on elementary students, where the students are still in the process ofdeveloping self-esteem, the results may be different.The third reason why our study did not find a direct effect of self-esteemon academic achievement may be due to the use of general (or global)self-esteem, as opposed to domain specific (or academic) self-esteem. Therehas been a great deal of research on academic self-esteem (or academicself-concept) and the construct has more consistently been found to affectacademic achievement. Rosenberg, Schooler, Schoenbach, and Rosenberg(1995), for example, found academic self-esteem had a significant impacton the school achievement of Grade 10 boys, whereas general self-esteemwas only found to affect their psychological well-being. Also, Marsh andYeung (1997) found that the academic self-concept of Grade 7–10 studentspositively predicted their subsequent academic achievement in Math,English, and Science, which in turn increased their later self-concept. Asa result, Marsh and Yeung advocated a reciprocal effects model in whichacademic self-concept and academic achievement concurrently affect eachother. Because general self-esteem has been found to consistently affectonly students’ psychological well-being, academic self-esteem would appearto be a more appropriate means of increasing their academic performanceas it has more consistently been found to affect academic achievement.Nevertheless, the more well-known construct of general self-esteem wasof greater interest and used in this study for two key reasons. First, ourliterature review on self-esteem showed that researchers in the past havecustomarily focused on general self-esteem instead of academic self-esteem(Bachman & O’Malley, 1986; Crocker & Luhtanen, 2003; Maruyama etal., 1981; Rosenberg et al., 1989; Skaalvik & Hagtvet, 1990). General selfesteemmay have been chosen in these past studies to test the commonbelief that it is beneficial in all areas of life, which is not supported as generalself-esteem is not consistently been found to predict academic achievement.The second reason was aptly stated by Morris Rosenberg, “...educatorsand policymakers almost invariably focus on precisely the wrongtype of self-esteem—global self-esteem—when they introduce interventionsto improve students performance in school” (Rosenberg et al., 1995, p. 154,italics added). Because general self-esteem is the construct commonly usedin educational interventions, it was of practical importance that this wasthe construct tested in our study to see how it compared to perceived control.We realize that comparing academic self-esteem and academic perceivedcontrol as predictors of college GPA may have been a more competitivecomparison as both are domain specific and have more consistently beenfound to influence academic achievement. 4 However, the objective of thisstudy was to elucidate an alternative construct to general self-esteem thatmore powerfully predicts the academic achievement of college students andcan be used in educational interventions on those students. The construct

320 ROBERT H. STUPNISKY ET AL.chosen for this study, based on logical, theoretical, and empirical bases(Perry, 1991; Perry et al., 2001; Perry, Hladkyj et al., 2005; Weiner, 1985,1995), was academic perceived control.4.2. PERCEIVED CONTROL IN COLLEGEPrior to our analyses, two explanations were offered as to why perceivedcontrol is important to college students’ academic achievement and ourresults lend support to both perspectives. First, regarding environmentalfactors, Perry (1991; Perry et al., 2001) argued that college can be alow-control environment for students due to the increased number of overwhelmingtasks that tend to make students feel out of control, which inturn has a negative effect on their academic performance. The significanteffect of perceived control on GPA supported this perspective because studentswho felt they had more influence over their environment tendedto succeed academically, whereas those who felt unable to control theircircumstances were more prone to failure.If students are more likely to experience academic success when they feelmore in control of their environment, an important question is what can bedone to increase the number of students who perceive college as a highlycontrollable environment? One approach would be for colleges to focus onincreasing the predictability and controllability of students’ academic tasks.For example, providing assistance with registration, choosing a major,selecting appropriate courses, applying for scholarships, securing financialresources, and finding appropriate housing, are all control-enhancingactivities that higher education institutions can focus on. Furthermore,professors can facilitate students’ perceptions of control by making theircourses more predictable. Detailed course syllabi, ample warning of testdates, suggested studying techniques, thorough explanations of class assignments,and increased availability for questions would all strengthenstudents’ perceptions of control and allow them to perform better academically.The second explanation for the impact of perceived control on collegestudents’ academic achievement stems from Weiner’s (1985, 1995) attributiontheory. Weiner posited that students who make uncontrollable attributionsfor their failures have lower perceptions of responsibility, increasedfeelings of shame, and decreased motivation, leading to diminished academicachievement. Alternatively, students who make controllable attributionsfor their failures have higher perceptions of responsibility, feel moreguilty, are more motivated, and subsequently attain higher grades. Wecontend that students with lower perceptions of control are more likelyto make uncontrollable attributions and will consequently perform morepoorly, whereas students with higher perceptions of control are more likelyto make controllable attributions and will have more academic success. Our

SELF-ESTEEM AND PERCEIVED CONTROL 321results supported this explanation, because students with higher, relative tolower, perceptions of control performed better academically. Consequently,it is important for first-year college students to have high perceptions ofcontrol so that they are more likely to make controllable attributions andmaintain an appropriate level of motivation to persevere through difficultsituations.This attributional explanation also lends support to an alternativemethod involving perceived control to improve college students’ academicperformances. Attributional retraining (AR) is a control-enhancing interventionbased on Weiner’s (1985) attribution theory that is designed tochange maladaptive attributions for poor performance into more adaptiveones (Perry, Hladkyj et al., 2005). More specifically, AR enhances achievementmotivation by encouraging students to attribute poor academic performanceto controllable factors, such as effort and strategy, instead ofuncontrollable factors, such as natural ability and professor quality (Hall,Perry, Ruthig, Hladkyj, & Chipperfield, 2006; Perry, Hetcher, Menec, &Weinberg, 1993; Ruthig, Perry, Hall, & Hladkyj, 2004). The positive impactof AR on college students’ motivation and academic achievement has beenwell replicated (for reviews see Forsterling, 1985; Perry et al., 1993; Perry,Hall et al., 2005), and the findings from this study further support the useof AR by suggesting students’ academic performances would be enhancedby encouraging students to view their poor performances in controllableways.4.3. STRENGTHS AND LIMITATIONSThe present study had several notable strengths that contributed to thevalidity of our findings. First, this was the first study to compare theimpact of self-esteem and perceived control on the academic achievementof college students. It should be noted that college students are becominga population of growing importance because the need for post-secondaryeducation is increasing (see Section 4.4). Second, our model utilized longitudinaldata obtained on three points in time: students’ high schoolgrades, their self-esteem and perceived control as they entered college, andtheir overall GPA at the end of their first year. This longitudinal designallowed for a more powerful test of the relationships between the variables.Third, we used standardized measures of both self-esteem and perceivedcontrol. The validated measure of perceived control was particularlyimportant given that many past studies have used proximal measuresof perceived control, such as locus of control (e.g., Ross & Broh, 2000),instead of using a specific scale. A perceived control scale allowed us tobetter understand the phenomenon of perceived control and to be moreconfident in the validity of our results. A fourth strength was that theresults were made more robust by statistically controlling for high school

322 ROBERT H. STUPNISKY ET AL.grades, gender, and age, all of which may affect self-esteem, perceived control,and/or GPA. Finally, we used “gold standard” measures of academicperformance, including students’ high school grades and GPAs, obtainedfrom institutional records.Of course, in field research involving longitudinal designs there arealmost always limitations and this study was no exception. First, the effectof perceived control on GPA, although statistically significant, was moderateto small in size and appears particularly small when compared to theeffect of high school grades on GPA. However, this result is important,considering that the effect accounts for students’ high school grades, age,and gender, making this study a more stringent test of perceived control’spredictive power. Furthermore, even small increases in GPA can make substantialdifferences to students’ success, including receiving a grade of Aversus B, being accepted to a particular faculty, being on the Dean’s list,or obtaining scholarships.A second limitation is that the sample used was somewhat specializedin terms of consisting entirely of college students enrolled in introductorypsychology. This specialized sample may have created biases as a result ofdisciplinary differences. Although the students were all enrolled in the samecourse, they came from a variety of academic disciplines (e.g., Arts = 27%,Science = 19%, and 43% unspecified or yet to decide on major), suggestingthere is diversity within the sample and makes the results generalizable toother college students. Finally, our study did not directly manipulate collegestudents’ self-esteem or perceived control. Consequently, causal statementscannot be made regarding the links between these constructs andacademic performance. It can be said, however, that perceived control measuredearly in the year was associated with higher academic achievement atthe end of the year and that self-esteem had no such association.4.4. IMPLICATIONS FOR ENHANCING COLLEGE STUDENTS’ ACADEMICSUCCESSAt the beginning of this article we made the presumption that most collegestudents desired to have high self-esteem and to be academically successful,but questioned whether self-esteem was the best construct to use as ameans of bolstering academic success. Our findings suggest that havinghigh self-esteem does not necessarily lead to higher grades for college students.Thus, how much time and effort should be invested in promotingself-esteem in higher education institutions given the increased importanceof obtaining a university education? Roberts, Clifton, Ferguson, Kampen,and Langlois (2005) provided a number of examples of how the valueof post-secondary education has increased over the past 40 years. Theyreported that enrollment in Canadian universities has increased by approximately5-fold (114,000 in 1961 vs. 580,000 in 1998), and the number

SELF-ESTEEM AND PERCEIVED CONTROL 323of university graduates with Bachelor’s degrees increased almost 6.5-fold(20,000 in 1961 vs. 126,000 in 1997). They also report that young peoplewith more education have, on average, lower unemployment rates andhigher incomes than young people with less education.As a consequence, it is in the best interest of both students and highereducation institutions that college students are successful academically.From a financial standpoint, failing a course or changing programs cancreate a financial burden on students by lengthening their time to graduationwhich adds thousands of dollars to their educational costs, not tomention future lost wages and possibly poorer career options. Similarly,when academic failure leads to withdrawal from the institution, lost tuitionrevenues for as few as 50 students can add up to hundreds of thousandsof dollars per year for the institution. As such, improving the academicperformance of college students is of significant practical and financialimportance to both students and postsecondary institutions.Our results suggest that perceived control has important consequencesfor students’ performance in college. Although we are not suggesting thatit is the sole contributor to academic performance, our results suggeststudents’ perceived control is more critical than their self-esteem. Basedon increased enrollment rates in college and the clear benefits of a postsecondaryeducation towards employment opportunities, students’ successin higher education is becoming increasingly important in today’s society.Therefore, the psychosocial constructs focused on as a means of reducingacademic failure among college students should be carefully chosen basedon empirical research.AcknowledgementsThis study was supported by a Social Science and Humanities ResearchCouncil of Canada (SSHRC) Canada graduate scholarship to the firstauthor, a SSHRC research grant #501-2002-0059 to the third author, aCanadian Institutes of Health Research (CIHR) postdoctoral fellowshipto the fourth author, and a SSHRC Canada graduate studentship to thefifth author. Special thanks to Lia Daniels, Steve Hladkyj, and NathanHall for feedback on an earlier version of the manuscript. Parts of thisresearch were presented at the annual meeting of the Canadian Society forthe Study of Education in Winnipeg, Manitoba, in April, 2004.Notes1 Students were recruited from introductory psychology classes in two different years, 1996(n=263) and 1998 (n=539). To test for differences between these two groups of students,the structural model was re-run using the year students participated in the study as agrouping variable (χ 2 = 625.22, df = 298, p

324 ROBERT H. STUPNISKY ET AL.test of invariance was conducted to compare the freely estimated model to a fully constrainedmodel across the groups (χ 2 = 638.66, df = 309). No statistical difference betweenthe groups was found (χ 2 = 13.44, df = 11, p>.05) and all paths from self-esteem andperceived control predicting GPA were virtually identical across the groups. As the twogroups model led to the same conclusions as the model in which the years were combined(see Section 3), the combined groups model was chosen to simplify presentation ofthe results.2 Academic achievement was also measured by students’ final grades in their introductorypsychology course. This measure of achievement may have been more strongly impactedby students’ self-esteem and perceived control because the students were recruited fromintroductory psychology classes for the study. As such, a structural model was analyzedwith final course grade in place of GPA. The model showed adequate fit to the data(χ 2 = 500.90, df = 149, p

SELF-ESTEEM AND PERCEIVED CONTROL 325Appendix A continuedItem M SDSelf-esteem1. I feel that I am a person of worth, at least on an equal plane 4.21 .85with others2. I feel that I have a number of good qualities. 4.25 .803. All in all, I’m inclined to feel that I am a failure. (R) 4.32 .904. I am able to do things as well as most other people. 3.92 .865. I feel I do not have much to be proud of. (R) 4.14 1.036. I take a positive attitude towards myself. 3.67 .997. On the whole, I am satisfied with myself. 3.57 1.008. I wish I could have more respect for myself. (R) 3.27 1.259. I certainly feel useless at times. (R) 3.22 1.2110. At times I think I am no good at all. (R) 3.64 1.24Perceived control1. I have a great deal of control over my academic performance 3.89 .94in my psychology course.2. The more effort I put into my courses, the better I do in 4.52 .78them.3. No matter what I do, I can’t seem to do well in my courses. 3.83 1.06(R)4. I see myself as largely responsible for my performance 4.45 .67throughout my college career.5. How well I do in my courses is often the ‘luck of the draw.’ 4.13 .84(R)6. There is little I can do about my performance in university. 4.53 .77(R)7. When I do poorly in a course, it’s usually because I haven’t 4.00 1.02given if my best effort.8. My grades are basically determined by things beyond my controland there is little I can do to change that. (R)4.31 .79Note. (R) indicates item was reverse coded.

326 ROBERT H. STUPNISKY ET AL.Appendix BVariances and covariances for all variables in study1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 221. Gender a .242. Age .01 .323. High school −.46 −.54 64.24. SE1 .03 .00 .62 .735. SE2 .04 .01 .20 .47 .636. SE3 .06 −.01 1.27 .34 .33 .817. SE4 .06 .02 .50 .32 .33 .26 .748. SE5 .03 −.00 .89 .39 .38 .47 .33 1.069. SE6 .06 −.01 .12 .40 .44 .43 .38 .45 .9910. SE7 .07 .01 .30 .39 .44 .42 .37 .46 .68 1.0011. SE8 .08 −.03 .12 .36 .42 .49 .32 .50 .57 .61 1.5612. SE9 .07 −.01 .49 .31 .31 .50 .32 .45 .64 .50 .76 1.4613. SE10 .06 .01 .32 .37 .40 .57 .36 .51 .64 .60 .83 1.05 1.5514. PC1 .03 −.01 1.57 .19 .15 .19 .15 .14 .11 .15 .14 .12 .15 .8815. PC2 −.01 −.03 .04 .17 .12 .11 .12 .13 .10 .08 .10 .06 .05 .18 .6016. PC3 .06 −.04 2.33 .20 .12 .29 .22 .27 .17 .19 .19 .31 .23 .33 .21 1.1217. PC4 −.01 −.00 .41 .15 .12 .10 .13 .09 .10 .11 .05 .09 .09 .22 .18 .17 .4418. PC5 .00 −.01 .88 .16 .11 .20 .11 .17 .13 .12 .19 .18 .17 .17 .17 .26 .16 .7019. PC6 .00 .01 .82 .12 .13 .13 .08 .14 .07 .09 .08 .08 .10 .22 .17 .24 .22 .23 .5920. PC7 .02 −.01 −.12 .11 .10 .08 .17 .04 .07 .07 .07 .05 .09 .27 .26 .17 .16 .14 .12 1.0521. PC8 .00 −.02 .76 .16 .13 .23 .10 .18 .10 .13 .15 .18 .20 .19 .18 .27 .19 .28 .26 .21 .6322. GPA −.03 .00 5.14 .09 .02 .17 .04 .12 .04 .10 .06 .13 .07 .23 .03 .32 .05 .11 .12 −.05 .11 .83aGender coded female = 1, male = 2.

SELF-ESTEEM AND PERCEIVED CONTROL 327ReferencesArbuckle, J.L. (2003). AMOS (Version 5.01) [Computer software]. Chicago, IL: SmallwatersCorporation.Bachman, J.G., & O’Malley, P.M. (1977). Self-esteem in young men: A longitudinalanalysis of the impact of educational and occupational attainment. Journal of Personalityand Social Psychology, 35(6), 365–380.Bachman, J.G., & O’Malley, P.M. (1986). Self-concepts, self-esteem, and educationalexperiences: The frog pond revisited (again). Journal of Personality and Social Psychology,50, 35–46.Baumeister, R.F., Campbell, J.D., Kruegger, J.I., & Vohs, K.D. (2003). Does high selfesteemcause better performance, interpersonal success, happiness, or healthier lifestyles?Psychological Science in the Public Interest, 4(1), 1–44.Bentler, P.M. (1990). Comparative fit indexes in structural equation models. PsychologicalBulletin, 107, 238–246.Browne, M.W., & Cudeck, R. (1993). Alternative ways of testing model fit. In K.A.Bollen & J.S. Long (Eds.), Testing structural equation models (pp. 445–455). NewburyPark, CA: Sage.Cappella, E., & Weinstein, R.S. (2001). Turning around reading achievement: Predictorsof high school students’ academic resilience. Journal of Educational Psychology, 93(4),758–771.Cassidy, S., & Eachus, P. (2000). Learning style, academic belief systems, self-report studentproficiency, and academic achievement in higher education. Educational Psychology:An International Journal of Experimental Educational Psychology, 20, 307–322.Chipperfield, J.G. (1993). Perceived barriers in coping with health problems: A twelveyearlongitudinal study of survival among elderly individuals. Journal of Aging andHealth, 5, 123–139.Clifton, R.A., Perry, R.P., Adams-Stubbs, C., & Roberts, L.W. (2004). Faculty environments,psychosocial dispositions, and the academic achievement of college students.Research in Higher Education, 45(8), 801–828.Crocker, J., & Luhtanen, R.K. (2003). Level of self-esteem and contingencies of selfworth:Unique effects on academic, social, and financial problems in college students.Personality and Social Psychology Bulletin, 29(6), 701–712.Demo, D.H., & Parker, K.D. (1988). Academic achievement and self-esteem amongblack and white college students. The Journal of Social Psychology, 127(4), 345–335.Diener, E., & Diener, M. (1995). Cross-cultural correlates of life satisfaction and selfesteem.Journal of Personality and Social Psychology, 68(4), 653–663.Findley, M.J., & Cooper, H.M. (1983). Locus of control and academic achievement: Aliterature review. Journal of Personality and Social Psychology, 44(2), 419–427.Forsterling, F. (1985). Attributional retraining: A review. Psychological Bulletin, 98, 495–512.Garber, J., & Seligman, M.E.P. (Eds.). (1980). Human helplessness: Theory and applications.New York, NY: Academic Press.Geraghty, M. (1996, July 19). More students quitting college before sophomore year,data show. The Chronicle of Higher Education, pp. A35–A36.Hall, N.C., Perry, R.P., Ruthig, J.C., Hladkyj, S., & Chipperfield, J.G. (2006). Primaryand secondary control in achievement settings: A longitudinal study of academicmotivation, emotions, and performance. Journal of Applied Social Psychology,36, 1430–1470.Haney, P., & Durlak, J.A. (1998). Changing self-esteem in children and adolescents: Ameta- analytic review. Journal of Clinical Child Psychology, 27(4), 423–433.

328 ROBERT H. STUPNISKY ET AL.Hu, L., & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structureanalysis: Conventional criteria versus new alternatives. Structural Equation Modeling,6(1), 1–55.Judge, T.A., Erez, A., Bono, J.E., & Thoresen, C.J. (2002). Are measures of self-esteem,neuroticism, locus of control, and generalized self-efficacy indicators of a commoncore construct? Journal of Personality and Social Psychology, 83(3), 693–710.Kalechstein, A.D., & Nowicki, S. Jr. (1997). A meta-analytic examination of the relationshipbetween control expectancies and academic achievement: An 11-year followupto Findley and Cooper. Genetic, Social, & General Psychology Monographs, 123(1),27–56.Kline, R.B. (2005). Principles and practice of structural equation modeling (2nd ed.). NewYork: Guilford Press.Lui, X., Kaplan, H.B., & Risser, W. (1992). Decomposing the reciprocal relationshipsbetween academic achievement and general self-esteem. Youth and Society, 24(2), 123–148.Marsh, H.W., Byrne, B.M., & Yeung, A.S. (1999). Causal ordering of academic self-conceptand achievement: Reanalysis of a pioneering study and revised recommendations.Educational Psychologist, 34(3), 155–167.Marsh, H.W., & Yeung, A.S. (1997). Causal effects of academic self-concept on academicachievement: Structural equation models of longitudinal data. Journal of EducationalPsychology, 89(1), 41–54.Maruyama, G., Rubin, R.A., & Kingsbury, G.G. (1981). Self-esteem and educationalachievement: Independent constructs with a common cause? Journal of Personality andSocial Psychology, 40, 962–975.Mooney, S.P., Sherman, M.F., & Lo Presto, C.T. (1991). Academic locus of control, selfesteem,and perceived distance from home as predictors of college adjustment. Journalof Counseling and Development, 69, 445–448.National Association for Self-esteem (NASE). Retrieved June 23, 2006 fromhttp://www.self-esteem-nase.org/boardmembers.shtml.Perry, R.P. (1991). Perceived control in college students: Implications for instruction inhigher education. In J. Smart (Ed.), Higher education: handbook of theory and research(Vol. 7, pp. 1–56). New York: Agathon Press.Perry, R.P. (2003). Perceived (academic) control and causal thinking in achievement settings.Canadian Psychologist, 44, 312–331.Perry, R.P., Hall, N.C., & Ruthig, J.C. (2005). Perceived (academic) control and scholasticattainment in higher education. In J. C. Smart (Ed.), Higher education: Handbookof theory and research (Vol. 20, pp. 363–436), The Netherlands: Springer.Perry, R.P., Hetcher, F.J., Menec, V.H., & Weinberg, L. (1993). Enhancing achievementmotivation and performance in college students: An attributional retraining perspective.Research in Higher Education, 34, 687–720.Perry, R.P., Hladkyj, S., Pekrun, R.H., Clifton, R.A., & Chipperfield, J.G. (2005).Perceived academic control and failure in college students: A three-year study ofscholastic attainment. Research in Higher Education, 46(5), 535–569.Perry, R.P., Hladkyj, S., Pekrun, R.H., & Pelletier, S.T. (2001). Academic control andaction control in the achievement of college students: A longitudinal field study. Journalof Educational Psychology, 93(4), 776–789.Pottebaum, S.M., Keith, T.Z., & Ehly, S.W. (1986). Is there a causal relation betweenself-concept and academic achievement? Journal of Educational Research, 79, 140–144.Roberts, L.W., Clifton, R.A., Ferguson, B., Kampen, K., & Langlois, S. (2005). Recentsocial trends in Canada, 1960–2000. Montreal, Quebec: McGill-Queen’s UniversityPress.

SELF-ESTEEM AND PERCEIVED CONTROL 329Robins, R.W., Hendin, H.M., & Trzesniewski, K.H. (2001). Measuring global selfesteem:Construct validation of a single-item measure and the Rosenberg Self-esteemScale. Personality and Social Psychology Bulletin, 27(2), 151–161.Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: PrincetonUniversity Press.Rosenberg, M., Schooler, C., & Schoenbach, C. (1989). Self-esteem and adolescent problems:Modeling reciprocal effects. American Sociological Review, 54, 1004–1018.Rosenberg, M., Schooler, C., Schoenbach, C., & Rosenberg, F. (1995). Global selfesteemand specific self-esteem: Different concepts, different outcomes. AmericanSociological Review, 60, 141–156.Ross, C.E., & Broh, B.A. (2000). The roles of self-esteem and the sense of personalcontrol in the academic achievement process. Sociology of Education, 73, 270–284.Rotter, J.B. (1966). Generalized expectancies for internal versus external control of reinforcement.Psychological Monographs, 80(1), 1–28.Ruthig, J.C., Haynes, T.L., Perry, R.P., & Chipperfield, J.G. (2007). Domain-specific optimisticbias: Implications for college students’ achievement and well-being. Social Psychologyof Education.Ruthig, J.C., Perry, R.P., Hall, N.C., & Hladkyj, S. (2004). Optimism and attributionalretraining: Longitudinal effects on academic achievement, test anxiety, and voluntarycourse withdrawal. Journal of Applied Social Psychology, 34, 709–730.Ruthig, J.C., Perry, R.P., Hladkyj, S., Hall, N.C., Pekrun, R., & Chipperfield, J.G.(2007). A longitudinal analysis of emotions and perceived control in an achievementsetting. Manuscript under review.Schönwetter, D., Perry, R.P., & Struthers, C.W. (1993). Students’ perceptions of controland success in the college classroom: Affects and achievement in different instructionconditions. Journal of Experimental Education, 61, 227–246.Scheirer, M.A., & Kraut, R.E. (1979). Increased educational achievement via self-conceptchange. Review of Educational Research, 49, 131–150.Skaalvik, E.M., & Hagtvet, K.A. (1990). Academic achievement and self-concept: Ananalysis of causal predominance in a developmental perspective. Journal of Personalityand Social Psychology, 58(2), 292–307.Skinner, E.A. (1996). A guide to constructs of control. Journal of Personality and SocialPsychology, 71(3), 549–570.Thompson, S.C. (1981). Will it hurt less if I can control it? A complex answer to asimple question. Psychological Bulletin, 90, 89–101.Tomas, J.M., & Oliver, A. (1999). Rosenberg’s Self-esteem Scale: Two factors or methodeffects. Structural Equation Modeling, 6(1), 84–98.Twenge, J.M., & Campbell, W.K. (2001). Age and birth cohort differences in self-esteem:A cross temporal meta-analysis. Personality and Social Psychology Review, 5(4), 321–344.Weiner, B. (1985). An attributional theory of achievement motivation and emotion.Psychological Review, 92, 186–200.Weiner, B. (1995). Judgements of responsibility: A foundation for a theory of social conduct.New York, NY: Guildford Press.Whiteside-Mansell, L., & Corwyn, R.F. (2003). Mean and covariance structures analyses:An examination of the Rosenberg Self-esteem Scale among adolescents andadults. Educational and Psychological Measurement, 63(1), 163–173.Woodard, P.G., & Suddick, D.E. (1992). Self-esteem of older adult college students’.Perceptual and Motor Skills, 74, 193–194.Zwick, R., & Sklar, J.C. (2005). Predicting college grades and degree completion usinghigh school grades and SAT scores: The role of student ethnicity and first language.American Educational Research Journal, 42(3), 439–464.

330 ROBERT H. STUPNISKY ET AL.Biographical notesRobert H. Stupnisky is a doctoral student in social psychology at the Universityof Manitoba. His research interests concern social cognition in academicdomains with a particular focus on the roles of perceived control,self-esteem, and attribution theory in the academic development and wellbeingof students.Office phone: +1-204-474-9338; Fax +1-204-474-7599; E-mail: umstupn1@cc.umanitoba.ca.Robert D. Renaud is an Assistant Professor of Education at the Universityof Manitoba. His research interests include performance indicators inhigher education, critical thinking, and structural equation modeling.Office phone: +1-204-474-6786; Fax: +1-204-474-7564; E-mail: renaudr@ms.umanitoba.ca.Raymond P. Perry is a Professor of Social and Health Psychology at theUniversity of Manitoba. His research focuses on motivation and socialcognition in achievement settings. Guided by social cognition theory, hestudies how psychological markers enable certain individuals to thrive inadverse conditions and how certain environments can thwart adaptation.Office phone: +1-204-474-9489; Fax: +1-204-474-7599; E-mail: rperry@ms.umanitoba.ca.Joelle C. Ruthig is an Assistant Professor of Psychology at the Universityof North Dakota. Her research mainly focuses on social cognition withinthe academic and health domains with a particular focus on the roles ofpsychosocial factors such as perceived control and optimism in achievementmotivation, physical health, and psychological well-being.Office phone: +1-701-777-3533; E-mail: joelle.ruthig@und.nodak.edu.Tara L. Haynes is a doctoral student in social psychology at the Universityof Manitoba. Her research interests are in social cognition with particularfocus on perceived control, causal attributions, and optimism.Office phone: +1-204-474-9338; Fax: +1-204-474-7599; E-mail: umhaynes@cc.umanitoba.ca.Rodney A. Clifton is a Professor in the Faculty of Education at the Universityof Manitoba, interested in sociology of classroom processes, academicmotivation, and achievement performance.Office phone: +1-204-474-9625; E-mail: clifton@ms.umanitoba.ca.

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTDOES HIGH SELF-ESTEEM CAUSE BETTERPERFORMANCE, INTERPERSONAL SUCCESS,HAPPINESS, OR HEALTHIER LIFESTYLES?Roy F. Baumeister, 1 Jennifer D. Campbell, 2 Joachim I. Krueger, 3 and Kathleen D. Vohs 41 Florida State University; 2 University of British Columbia, Vancouver, British Columbia, Canada; 3 Brown University;and 4 University of UtahSummary—Self-esteem has become a household word.Teachers, parents, therapists, and others have focused effortson boosting self-esteem, on the assumption that high selfesteemwill cause many positive outcomes and benefits—anassumption that is critically evaluated in this review.Appraisal of the effects of self-esteem is complicated byseveral factors. Because many people with high self-esteemexaggerate their successes and good traits, we emphasize objectivemeasures of outcomes. High self-esteem is also a heterogeneouscategory, encompassing people who frankly accepttheir good qualities along with narcissistic, defensive, andconceited individuals.The modest correlations between self-esteem and schoolperformance do not indicate that high self-esteem leads togood performance. Instead, high self-esteem is partly the resultof good school performance. Efforts to boost the selfesteemof pupils have not been shown to improve academicperformance and may sometimes be counterproductive. Jobperformance in adults is sometimes related to self-esteem, althoughthe correlations vary widely, and the direction of causalityhas not been established. Occupational success mayboost self-esteem rather than the reverse. Alternatively, selfesteemmay be helpful only in some job contexts. Laboratorystudies have generally failed to find that self-esteem causesgood task performance, with the important exception thathigh self-esteem facilitates persistence after failure.People high in self-esteem claim to be more likable and attractive,to have better relationships, and to make better impressionson others than people with low self-esteem, butobjective measures disconfirm most of these beliefs. Narcissistsare charming at first but tend to alienate others eventually.Self-esteem has not been shown to predict the quality orduration of relationships.High self-esteem makes people more willing to speak up ingroups and to criticize the group’s approach. Leadershipdoes not stem directly from self-esteem, but self-esteem mayhave indirect effects. Relative to people with low self-esteem,Address correspondence to Roy F. Baumeister, Department of Psychology,Florida State University, Tallahassee, FL 32306-1270.those with high self-esteem show stronger in-group favoritism,which may increase prejudice and discrimination.Neither high nor low self-esteem is a direct cause of violence.Narcissism leads to increased aggression in retaliationfor wounded pride. Low self-esteem may contribute to externalizingbehavior and delinquency, although some studies havefound that there are no effects or that the effect of self-esteemvanishes when other variables are controlled. The highest andlowest rates of cheating and bullying are found in different subcategoriesof high self-esteem.Self-esteem has a strong relation to happiness. Although theresearch has not clearly established causation, we are persuadedthat high self-esteem does lead to greater happiness.Low self-esteem is more likely than high to lead to depressionunder some circumstances. Some studies support the buffer hypothesis,which is that high self-esteem mitigates the effects ofstress, but other studies come to the opposite conclusion, indicatingthat the negative effects of low self-esteem are mainlyfelt in good times. Still others find that high self-esteem leads tohappier outcomes regardless of stress or other circumstances.High self-esteem does not prevent children from smoking,drinking, taking drugs, or engaging in early sex. If anything,high self-esteem fosters experimentation, which may increaseearly sexual activity or drinking, but in general effects of selfesteemare negligible. One important exception is that highself-esteem reduces the chances of bulimia in females.Overall, the benefits of high self-esteem fall into two categories:enhanced initiative and pleasant feelings. We have notfound evidence that boosting self-esteem (by therapeutic interventionsor school programs) causes benefits. Our findingsdo not support continued widespread efforts to boost selfesteemin the hope that it will by itself foster improved outcomes.In view of the heterogeneity of high self-esteem, indiscriminatepraise might just as easily promote narcissism, withits less desirable consequences. Instead, we recommend usingpraise to boost self-esteem as a reward for socially desirablebehavior and self-improvement.Most people feel that self-esteem is important. It is difficult,if not impossible, for people to remain indifferent to informationthat bears on their own self-esteem, such as being told thatthey are incompetent, attractive, untrustworthy, or lovable. Increasesand decreases in self-esteem generally bring strongemotional reactions. Moreover, these fluctuations are often coincidentwith major successes and failures in life. SubjectiveVOL. 4, NO. 1, MAY 2003 Copyright © 2003 American Psychological Society 1

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemexperience creates the impression that self-esteem rises whenone wins a contest, garners an award, solves a problem, or gainsacceptance to a social group, and that it falls with correspondingfailures. This pervasive correlation may well strengthen the impressionthat one’s level of self-esteem is not just the outcome,but indeed the cause, of life’s major successes and failures.But is self-esteem a cause of important consequences inlife? In this monograph, we report the results of a survey ofmajor research findings bearing on this question. Our missionwas to conduct a thorough review of empirical findings—emphasizingthe most methodologically rigorous research studies—toascertain whether high self-esteem is in fact a cause ofpositive or negative outcomes. We anticipated we would findthat self-esteem has positive value for bringing about some hypothesizedbenefits, but not others. Such a pattern would presumablyallow an accurate and nuanced understanding of justwhat high self-esteem is good for. This would be beneficialboth for theory (in that it would promote a better understandingof self-esteem as well as the outcomes it predicts) and for practicalapplications—and even for determining whether efforts atboosting self-esteem are worth undertaking in order to solveparticular social problems.Self-esteem is literally defined by how much value peopleplace on themselves. It is the evaluative component of selfknowledge.High self-esteem refers to a highly favorable globalevaluation of the self. Low self-esteem, by definition, refersto an unfavorable definition of the self. (Whether this signifiesan absolutely unfavorable or relatively unfavorable evaluationis a problematic distinction, which we discuss later in connectionwith the distribution of self-esteem scores.) Self-esteemdoes not carry any definitional requirement of accuracy whatsoever.Thus, high self-esteem may refer to an accurate, justified,balanced appreciation of one’s worth as a person andone’s successes and competencies, but it can also refer to an inflated,arrogant, grandiose, unwarranted sense of conceited superiorityover others. By the same token, low self-esteem canbe either an accurate, well-founded understanding of one’sshortcomings as a person or a distorted, even pathologicalsense of insecurity and inferiority.Self-esteem is thus perception rather than reality. It refers toa person’s belief about whether he or she is intelligent and attractive,for example, and it does not necessarily say anythingabout whether the person actually is intelligent and attractive.To show that self-esteem is itself important, then, researchwould have to demonstrate that people’s beliefs about themselveshave important consequences regardless of what the underlyingrealities are. Put more simply, there would have to bebenefits that derive from believing that one is intelligent, regardlessof whether one actually is intelligent. To say this is notto dismiss self-esteem as trivial. People’s beliefs shape their actionsin many important ways, and these actions in turn shapetheir social reality and the social realities of the people aroundthem. The classic study Pygmalion in the Classroom, byRosenthal and Jacobson (1968), showed that teachers’ false,unfounded beliefs about their students later became objective,verifiable realities in the performance of those students. In thesame way, it is quite plausible that either high or low selfesteem,even if initially false, may generate a self-fulfillingprophecy and bring about changes in the objective reality of theself and its world.Then again, self-esteem might not bring about such changes.Many researchers, clinicians, teachers, parents, and punditshave taken it as an article of faith that high self-esteem willbring about positive outcomes. Such an assumption was perhapsreasonable several decades ago, given the lack of firmdata either way and the anecdotal impressions and theoreticalbases for assuming that self-esteem has strong effects. It is particularlyunderstandable that practitioners would accept this assumptionwithout proof, because they cannot generally affordto admonish their suffering clients to hang on for a few decadesuntil needed research is conducted. They must use the best evidenceavailable at the time to design their interventions.By now, however, the excuse of inadequate data is beginningto wear thin. The fascination with self-esteem that beganto spread during the 1970s infected researchers too, and in thepast couple of decades, a number of methodologically rigorous,large-scale investigations on the possible effects of selfesteemhave been conducted. We do not think all the final answersare in, but many of them are taking shape. There is nolonger any justification for simply relying on anecdotes, impressions,and untested assumptions about the value of selfesteem.WHY STUDY SELF-ESTEEM?In the heady days of the 1970s, it might have seemed possibleto assert that self-esteem has a causal effect on every aspectof human life, and by the 1980s, the California legislaturemight well have been persuaded that funding a task force to increasethe self-esteem of Californians would ultimately producea huge financial return because reducing welfare dependency,unwanted pregnancy, school failure, crime, drug addiction, andother problems would save large amounts of taxpayers’ money.However, as Karl Marx, Sigmund Freud, and other grandthinkers could assert if they were alive today, even the mostelaborate and persuasive theories about human behavior do notgenerally receive empirical support in all aspects. Thus, wenote at the outset that we did not expect all the extravagantclaims of the self-esteem movement to be supported.Even if the self-esteem movement was wrong in crucial respects,its positive aspects and contributions deserve to be recognizedand celebrated. The self-esteem movement showedthat the American public was willing to listen to psychologistsand to change its institutional practices on the basis of whatpsychology had to teach. It would not be in psychology’s bestinterest to chastise the American public for accepting the ad-2 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.vice of psychologists. If errors were committed, perhaps psychologistsshould reduce their own self-esteem a bit andhumbly resolve that next time they will wait for a more thoroughand solid empirical basis before making policy recommendationsto the American public. Regardless of the outcomeof the self-esteem movement, it showed that there is a voice forpsychology in public policy and discourse. If psychology usesthat voice judiciously, it may still be able to make a major contributionto the well-being of society.The Appeal of Self-EsteemAs self-aware and self-reflective creatures, many people intuitivelyrecognize the importance of self-esteem. Not surprisingly,a great deal of psychological theorizing has focused onthe motivation to protect and, if possible, enhance self-esteem.Research is showing that even psychodynamic defense mechanisms,which Freud originally understood as ways of keepingthreatening sexual and aggressive impulses at bay, serve asstrategies to bolster self-esteem (for a review, see Baumeister,Dale, & Sommer, 1998).But the desire to feel good about oneself is certainly not theonly self-related motive at play. Having to cope with reality,people are also motivated to perceive themselves accuratelyand admit awareness of their undesirable characteristics (Swann,Stein-Seroussi, & Giesler, 1992; Trope, 1986). Nevertheless,people would rather learn positive things about themselvesthan negative things (Sedikides, 1993). Although they maywant to know whether they are good or not, they much preferto learn that they are good.Over the past few decades, the need for high self-esteem hasrisen from an individual to a societal concern. North Americansociety in particular has come to embrace the idea that highself-esteem is not only desirable in its own right, but also thecentral psychological source from which all manner of positivebehaviors and outcomes spring. This strong psychologicalclaim has begun to permeate popular beliefs. Its corollary, theidea that low self-esteem lies at the root of individual and thussocietal problems and dysfunctions, has obvious implicationsfor interventions on both the individual and the societal level.The hope that such interventions might work has sustained anambitious social movement. Nathaniel Branden, a leading figurein the self-esteem movement, stated categorically that“self-esteem has profound consequences for every aspect ofour existence” (Branden, 1994, p. 5), and, more pointedly, thathe “cannot think of a single psychological problem—from anxietyand depression, to fear of intimacy or of success, to spousebattery or child molestation—that is not traceable to the problemof low self-esteem” (Branden, 1984, p. 12). Other advocatesof the movement have endorsed this sentiment. AndrewMecca, for example, is cited as saying that “virtually every socialproblem can be traced to people’s lack of self-love”(Davis, 1988, p. 10).Academic and professional psychologists have been morehesitant to endorse strong categorical claims. Eminent clinicalpsychologist Albert Ellis, for example, is convinced that “selfesteemis the greatest sickness known to man or woman becauseit’s conditional” (cited in Epstein, 2001, p. 72). Accordingto Ellis, people would be better off if they stopped trying toconvince themselves that they are worthy. Others believe thatconcerns about self-esteem are a peculiar feature of Western individualistcultures. According to this perspective, the searchfor high self-esteem is not a universal human motive, but a culturalor ideological artifact. Indeed, such a motive is difficult todetect in collectivist cultures, and especially in Japan (Heine,Lehman, Markus, & Kitayama, 1999). Even in Western culture,the need for high self-esteem seems to be a rather recent development(Baumeister, 1987). The Judeo-Christian tradition haslong considered modesty and humility as virtues conducive tospiritual growth. In this tradition, high self-esteem is suspectbecause it opens the door to sentiments of self-importance.Medieval theologians considered pride or vainglory to be particularlysatanic and thus a deadly sin. To combat it, religiousdevotees cultivated an unattractive appearance (e.g., shorn hair,no makeup, unfashionable clothes, no jewelry), spoke withself-effacement, and submitted to degrading exercises (e.g.,begging, prostrations, self-flagellations).Such practices are but a faint memory in contemporary popularculture, in which high self-esteem seems to reign supreme.Prodded by Assemblyman John Vasconcellos, the then governorof California George Deukmeijian agreed in 1986 to fund aTask Force on Self-Esteem and Personal and Social Responsibilitywith a budget of $245,000 per annum for several years.Vasconcellos argued that raising self-esteem would help solvemany of the state’s problems, including crime, teen pregnancy,drug abuse, school underachievement, and pollution. At onepoint, he expressed the hope that raising self-esteem wouldhelp balance the state’s budget because people with high selfesteemearn more money than people with low self-esteem andtherefore pay more taxes (Winegar, 1990). It is easy to dismissand satirize such claims (Dawes, 1994). However, Vasconcellosand the task force also speculated astutely about the possibilitythat self-esteem might protect people from being overwhelmedby life’s challenges and thus reduce failures and misbehaviors,much as a vaccine protects against disease.Concurrent with its activities in the field, which includedcreating self-esteem committees in many California counties,the task force assembled a team of scholars to survey the relevantliterature. The results were presented in an edited volume(Mecca, Smelser, & Vasconcellos, 1989). Echoing Branden(1984), Smelser (1989) prefaced the report by stating that“many, if not most, of the major problems plaguing societyhave roots in the low self-esteem of many of the people whomake up society” (p. 1). But the findings did not validate thehigh hopes of the task force, and Smelser had to acknowledgethat “one of the disappointing aspects of every [italics added]VOL. 4, NO. 1, MAY 2003 3

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemchapter in this volume . . . is how low the associations betweenself-esteem and its [presumed] consequences are in research todate” (p. 15). Given that the correlations were so low, the questionof whether low self-esteem in fact caused the societalproblems did not even arise.The lack of supportive data created a dilemma. Should a notionas attractive as self-esteem be abandoned and replacedwith more promising concepts, or should the validity of the evidencebe questioned? The editors and the authors opted for amix of these two strategies. Some retreated to a defense of selfesteemon a priori grounds. Undeterred, Smelser (1989) maintainedthatdiminished self-esteem stands as a powerful independent variable(condition, cause, factor) in the genesis of major social problems. Weall know this to be true, and it is really not necessary to create a specialCalifornia task force on the subject to convince us. The real problemwe must address—and which the contributors to this volume address—ishow we can determine that it is scientifically true. (p. 8)Others, however, acknowledged the limitations of the findingsand called for additional study, or tried to fit more complex theoreticalmodels of self-knowledge to the data. Our report is focusedprimarily on studies conducted since the review by theCalifornia task force. Instead of examining the merits of themore complex models of self, we have retained the hypothesisthat global self-esteem causes desirable, adaptive, and beneficialbehaviors. There is a certain beauty to this hypothesis becauseit is simple, clear, and testable. There have also beensufficient methodological advances in study design and statisticalanalysis that warrant a fresh look at the evidence.Meanwhile, the self-esteem movement was not deterred bythe disappointing findings of the task force. After it was disbandedin 1995, the National Council for Self-Esteem inheritedits mandate, which was subsequently taken on by the NationalAssociation for Self-Esteem, or NASE. Vasconcellos (now amember of the California Senate) and Jack Canfield (ChickenSoup for the Soul) are on NASE’s advisory board, and suchmedia personalities as Anthony Robbins (Unlimited Power),Bernie Siegel (Love, Medicine, and Miracles), and GloriaSteinem (A Revolution From Within: A Book of Self-Esteem)are members of a “Masters Coalition,” created by NASE. Themission statement of NASE minces no words about the presumedbenefits of self-esteem. Its goal is to “promote awarenessof and provide vision, leadership and advocacy forimproving the human condition through the enhancement ofself-esteem” (NASE, 2000). The goal of the Masters Coalitionis no less ambitious. “It is hoped that the Master Coalition can,in a meaningful way, facilitate the actualization of society andlead to the amelioration, if not elimination, of various negativeinfluences which have operated in part to trivialize and demeanthe human condition” (NASE, 2000).It is hard not to conclude that the self-esteem movement hasignored its own major scholarly document (i.e., the Mecca etal., 1989, volume). In the quest for enhanced self-esteem, anytool in the psychological—and pseudopsychological—box isthrown into the fray, includingdisparate psychological models that have given rise to such popularnotions as the “inner child”; the “self-image”; principles of propergrieving; “super learning”; “community networking”; “relaxationtechniques” and their effects on overall mental and physical well-being;the principles of “neuro-linguistic programming”; and the wellfoundedscientific basis for the connection between the body and themind and the effect of this interface on overall wellness. (NASE,2000)Even a contributor to the volume edited by Mecca et al. (1989)argued that self-esteem must be enhanced, although its causalrole is far from established. “To abandon the search for esteemrelatedsolutions . . . is to admit defeat before exploring all ouroptions” (Covington, 1989, p. 74).Was it reasonable to start boosting self-esteem before all thedata were in? Perhaps. We recognize that many practitionersand applied psychologists must deal with problems before allthe relevant research can be conducted. Still, by now there areample data on self-esteem. Our task in this monograph is totake a fresh look and provide an integrative summary.An Epidemic of Low Self-Esteem?A key assumption of the self-esteem movement is that toomany people have low self-esteem. Under this assumption,raising self-esteem becomes a meaningful goal. But what does“too many” mean? Self-esteem scales are designed to capturevalid individual differences that exist in a population. Thus, agood measure will yield a distribution of scores from low tohigh. However, unlike some other measurement instruments,such as IQ tests, that are constructed to yield symmetrical distributionscentered around an arbitrary mean (e.g., 100), selfesteemscales allow skewed distributions to emerge. Theaverage score typically lies far above the midpoint of the scale,often by more than a standard deviation (Baumeister, Tice, &Hutton, 1989). The fact that most people score toward the highend of self-esteem measures casts serious doubt on the notionthat American society is suffering from widespread low selfesteem.If anything, self-esteem in America is high. The averageperson regards himself or herself as above average.The skewed distribution of self-esteem scores raises twomethodological issues. First, when researchers split samples atthe median to distinguish between respondents with high versuslow self-esteem, the range of scores among respondentsclassified as having low self-esteem is much greater than therange of scores among respondents classified as having highself-esteem. A good number of respondents in the low self-esteemcategory have scores above the midpoint of the scale. Inother words, the classification of a person as someone with lowself-esteem has no longer an absolute, but only a relative meaning.Second, correlations involving variables with skewed dis-4 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.tributions tend to be smaller than correlations involving variableswith symmetric distributions. Moreover, when self-esteem israised selectively for those respondents with the lowest initialvalues, correlations between self-esteem and relevant outcomevariables shrink further, not necessarily because the elevationof self-esteem had the desired causal effect, but simply becauseof the restriction in the range of scores. It is always necessaryto ask whether relevant outcomes also changed in the desireddirection.The standard finding that most self-esteem scores are highraises the possibility that at least some scores are affected bydeliberate or unwitting self-enhancement (Krueger, 1998).Brown (1986), for example, found that people high in self-esteemwere also most likely to rate themselves more positivelythan they rated other people. Because self-enhancement mayinvolve invalid and undesirable distortions of the self-concept,it is unwarranted to rush to boost everyone’s self-esteem.In short, we find no evidence that modern Western societiesare suffering from an epidemic of low self-esteem. If anything,self-esteem seems generally high in most North American samples.Regardless of their race, gender, or socioeconomic status,Americans already appear to live in a “culture of self-worth”(Twenge & Campbell, 2001, p. 325). Indeed, levels of selfesteemincreased at a time when the self-esteem movement bemoanedthe lack of self-love. Disturbingly, academic performancedecreased at the same time (Twenge & Campbell, 2001).PROBLEMS AND CHALLENGESMeasurement of Self-EsteemMany scales are available for measuring self-esteem, anddifferent investigations have used different ones, which compoundsthe difficulty of comparing results from different investigations(especially if the results are inconsistent). Blascovichand Tomaka (1991) reviewed multiple measures and foundthem of uneven quality, giving high marks to only a few (suchas Fleming & Courtney’s, 1984, revision of Janis & Field’s,1959, scale, and Rosenberg’s, 1965, global self-esteem measure).In essence, self-esteem scales ask people to rate themselvesin response to questions such as “Are you a worthwhileindividual?” “Are you good at school or work?” “Do peoplelike you?” and “Are you reliable and trustworthy?” When researcherscheck self-esteem measures against the so-called liescales (also called measures of social desirability, because theyassess tendencies to give distorted, even unrealistic answersjust to make a good impression), they conclude that self-esteemscores are somewhat contaminated by people’s efforts to makethemselves look good. These measures also obscure neededdistinctions between defensive, inflated, narcissistic, and socalledgenuine high self-esteem. (We discuss different varietiesof high self-esteem in the next section.) Unfortunately, there isno objective criterion against which to compare self-reportedself-esteem, because of the nature of the construct: Self-esteemessentially consists of how a person thinks about and evaluatesthe self. In the case of intelligence, for example, self-ratingscan be compared against objective performance on intellectualtests, and the results can (and often do) show that people’s selfreportsof their own intelligence are wrong. But there is noknown basis for saying that certain people really have more orless self-esteem than they think they have.To overcome these measurement problems, some researchersmeasure implicit, or unfakeable, self-esteem by using a varietyof subtle methods, such as reaction times to good and badthoughts that can be paired with the self (Greenwald & Farnham,2000). Though promising, this research has only recentlybegun, and it therefore does not play a significant role in thisreview. Despite the potential pitfalls of explicit (i.e., self-report)measures, the fact that scores on different scales are positivelycorrelated (e.g., Greenwald & Farnham, 2000) is an indicationthat they can be used with some confidence. Even more significantly,the Rosenberg scale, which is by far the most popularamong researchers, has been shown to be highly reliable (e.g.,if a person completes the scale on two occasions, the twoscores tend to be similar). As a measure of global self-esteem,this scale is unidimensional (Gray-Little, Williams, & Hancock,1997; Robins, Hendin, & Trzesniewski, 2001). Indeed,its reliability is so high that a single item (“I have high selfesteem”)may be sufficient (Robins et al., 2001).Heterogeneity of High Self-EsteemThe high internal consistency of self-esteem measures maymask the possibility that a variety of psychological processescontribute to high (or low) scores. One approach to studyingthe heterogeneity of self-esteem is to examine the pattern ofscores across multiple measurement instruments. Schneiderand Turkat (1975) suspected that some people’s high self-esteemis defensive rather than genuine, and that these individualscould be identified if they also scored high on the Marlowe-Crowne Social Desirability Scale (Crowne & Marlowe, 1960).The concept of defensive self-esteem has recently been refinedby the distinction between deception of others (i.e., impressionmanagement) and deception of the self (see Paulhus, 2002, fora review). High self-esteem is considered defensive if it is coupledwith high scores on a self-deception scale (which hasitems such as “I always know why I do things”).Taking a different approach, Kernis and his colleagues (seeKernis & Waschull, 1995, for a review) measure both the overalllevel and the temporal stability of self-esteem. In manystudies, the stability of self-esteem, either by itself or in combinationwith level of self-esteem, has been shown to predict behavioraloutcomes. Baumeister and his colleagues (e.g., Baumeister,1993; Baumeister, Smart, & Boden, 1996) found that behaviorsand outcomes are often more variable for people high in self-VOL. 4, NO. 1, MAY 2003 5

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemesteem than for people low in self-esteem. For example, studiesoften fail to yield a significant correlation between aggressionor violence and self-esteem, in part because high self-esteem isassociated with both the presence and the absence of aggression.Kernis’s work leads to similar conclusions. People withhigh but unstable self-esteem score higher on measures of hostilitythan do people with low self-esteem (whether stable orunstable), whereas people with high but stable self-esteem arethe least hostile (Kernis, Grannemann, & Barclay, 1989). Inother words, people with high self-esteem are found at the extremesof both high and low hostility.A third approach is to measure narcissism along with, oreven instead of, self-esteem. The construct of narcissism involveshighly favorable, even grandiose views of self, a senseof being special or unique, fantasies of personal brilliance orbeauty, and the belief that one is entitled to privileges and admirationby others (see American Psychiatric Association,2000). In normal populations, scores on the Narcissistic PersonalityInventory (Emmons, 1984; Raskin & Hall, 1981) correlatesubstantially with self-esteem (Rhodewalt & Morf, 1995;Sinha & Krueger, 1998). The heterogeneity of high self-esteemis indicated by the finding that some people with high selfesteemare not narcissistic, whereas others are; the reverse isnot true (i.e., there are very few narcissists with low selfesteem).Bushman and Baumeister (1998) measured both selfesteemand narcissism in laboratory studies of aggression. Narcissism,but not self-esteem, predicted aggression, and onlywhen the target was someone who had previously insulted theparticipant. The idea that narcissistically high self-esteem isunhealthy is further supported by the finding that narcissists’high self-esteem tends to be unstable (Rhodewalt, Madrian, &Cheney, 1998) and self-defensive (Paulhus, 1998).Taken together, these lines of research suggest that somelow correlations between self-esteem and socially relevant behaviorsor outcomes conceal the tendency for different categoriesof high self-esteem to produce opposite responses. Researcherswho believe in the value of so-called genuine or healthy selfesteemmay find that they can obtain more valid correlationswith desirable outcomes if they control variables such as narcissism,self-deception, or temporal stability. The broader implicationis, however, that the category of people with highself-esteem is a mixed bag that contains individuals whose selfopinionsdiffer in important ways.Theorists and researchers have linked low self-esteem toother constructs, generally focusing on the links between lowself-esteem and pathologies such as depression. Recently, lowself-esteem has begun to be associated with more general conceptssuch as emotional lability (a tendency for strong emotionsto occur in both directions) and low internal locus of control (ageneralized belief that the self is not in control of what happens).Although there are moderate associations between low selfesteemand pathology (see the section on Coping and Depression),there is also evidence against the notion that the constructsof low self-esteem and depression are isomorphic.Correlations between self-esteem and depression are of onlymoderate strength, ranging from .4 to .6 (e.g., Joiner, Alfano, &Metalsky, 1992). Also, a theoretical standpoint indicates thatalthough the psychological processes associated with selfesteemand depression may overlap somewhat (e.g., Abramson,Metalsky, & Alloy, 1989), they are not identical. Rather, theconcept of depression has been characterized by a constellationof symptoms, of which low self-esteem is one (Roberts &Monroe, 1999). But low self-esteem is neither necessary norsufficient for depression.A recent analysis of the interrelations among self-esteem,neuroticism, locus of control, and feelings of being effectivepoints to a more serious methodological problem regarding theiroverlap. Judge, Erez, and Bono (2002) found that these constructsare highly related and reflect one overarching construct(which they said was neuroticism, broadly defined). Judge et al.concluded that the ability of any of these constructs to uniquely(i.e., independently) predict outcomes is quite poor, and theyurged psychological scientists to begin thinking of each constructin tandem with the others. We concur with this suggestionand hope that researchers will include more of these measures instudies of the effect of self-esteem on objective outcomes.Global Versus Domain-Specific Self-EsteemThe heterogeneous nature of people who score high on selfesteemmeasures is not the only reason why the predictivepower of global measures of self-esteem is limited. Anotherreason is that it is difficult to detect a correspondence betweena global attitude and specific behaviors or outcomes (Fishbein& Ajzen, 1975). It is not to be expected, for example, that aglobal sense of being worthy, competent, and popular will predictperformance on a trigonometry quiz. Many people mayconsider mathematical ability to be irrelevant to global selfappraisal,and so their self-esteem could be utterly irrelevant tohow well they can perform numerical calculations. Predictionsimprove when self-esteem is measured for the domain of interestand among people who consider this domain to be personallyimportant (Crocker & Wolfe, 2001). If relevant domainsare hierarchically organized, it is important to measure selfesteemat the appropriate level of specificity. If the domain becomestoo narrow, the assessment of specific self-esteem mayyield only trivial results. For example, a high correlation betweenpeople’s success at doing long division and their self-evaluationfor this task may simply result from people’s awareness of theirability in this domain. If so, any attempts to improve performanceby way of enhancing self-esteem would fail.The difficulties of relating global self-esteem to specific behaviorsmay be overcome, in part, by aggregating behaviorsinto bundles. But aside from differences in levels of specificityof measurement, there is also a difference in affectivity. Globalself-esteem is heavily invested with feelings about the self,whereas specific facets of self-esteem include a variety of selfrelatedthoughts (Brown, 1998; Rosenberg, Schooler, Schoen-6 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.bach, & Rosenberg, 1995). It is perhaps because of its high affectiverelevance that global self-esteem has been the preferredtarget for well-intentioned interventions. Interventions at thelevel of domain-specific self-esteem would arguably becomeboth more fragmented and less affectively charged. For thesereasons, we focus our review on global self-esteem.FloccinaucinihilipilificationThe word floccinaucinihilipilification is allegedly the longestword in the Oxford English Dictionary, and it is defined as “theaction or habit of estimating as worthless.” The definition of lowself-esteem involves making a disparaging or low-worth judgmentabout the self, yet there is mounting evidence that peoplewith low self-esteem are not merely negative about themselves.Rather, they express a generally negative attitude toward manyevents, circumstances, people, and other realities. That is, theyhave a tendency toward floccinaucinihilipilification.The problem this raises for the researcher is twofold. First,it is difficult to distinguish the general negativity from the specificlow self-esteem. A good example of this problem emergedin the research literature on prejudice. In early work, researchersassumed that people with low self-esteem would have moreprejudice than others toward out-groups, and studies in whichpeople simply rated other groups seemingly confirmed thatpeople with low self-esteem expressed more negative attitudestoward them than did people with higher self-esteem (seeWills, 1981, for review). But Crocker and Schwartz (1985)pointed out that if people rate themselves and their in-groupsnegatively, it is hardly fair to label them as prejudiced for ratingout-groups negatively, too. Such general floccinaucinihilipilificationcan be corrected by subtracting the ratings of self or thein-group from the ratings of the out-group, and when this wasdone, the finding was reversed: People with high self-esteemshowed greater in-group bias and greater prejudice than peoplewith low self-esteem (see also Aberson, Healy, & Romero,2000; Crocker & Luhtanen, 1990; Crocker, Thompson, McGraw,& Ingerman, 1987).Second, floccinaucinihilipilification raises the very substantialdanger that self-reports will show spurious correlations betweenself-esteem and other outcomes. A general tendencytoward floccinaucinihilipilification may cause certain people todescribe themselves in disparaging terms, thus generating lowscores on self-esteem, as well as to describe their lives and outcomesin disparaging terms, thus furnishing the appearance ofmaladaptive behaviors and pathology. This is one reason why itis important to obtain objective outcome measures.Need for Objective MeasuresAll too often, results involving self-esteem are based on theself-report of outcome variables. The problem this raises canbe illustrated by the case of a researcher who asks people,“How good are you at juggling red beanbags?” and then takestheir answers as indications of their self-esteem in the beanbagjugglingsphere. If the researcher also decides to use self-reportsto measure performance, almost exactly the same questioncould be used as a measure of performance. The resulting highcorrelation between the measures of self-esteem and performancewould reflect nothing more than the fact that the samequestion was asked to measure different constructs.It is essential to keep in mind that self-esteem is measuredalmost exclusively (and unavoidably) by self-report. Peoplescore high in self-esteem because they respond to a questionnaireby endorsing favorable statements about themselves. Thehabit of speaking well of oneself does not abruptly cease whenthe respondent turns from the self-esteem scale to the questionnaireasking for self-report of other behaviors. People who liketo describe themselves in glowing terms will be inclined to reportthat they get along well with others, are physically attractive,do well in school and work, refrain from undesirableactions, and the like. That is how they get high scores in selfesteem,but researchers may easily mistake this identical tendencyas evidence that self-esteem predicts or even causes abroad range of positive outcomes. Over and over during oursurvey of the literature, we found that researchers obtainedmore impressive evidence of the benefits of self-esteem whenthey relied on self-reported outcomes than when they relied onobjective outcomes. The research on the relationship betweenself-esteem and physical attractiveness provides a good exampleof such a discrepancy. Most self-esteem scales do not containitems that specifically ask whether respondents considerthemselves physically attractive, but they do measure the globaltendency to speak well of oneself. It seems plausible thatpeople who speak well of themselves in general would ratetheir physical attractiveness more highly than others. Thenagain, it is plausible that physically attractive people would endup with higher self-esteem than other people, if only becauseattractive people are treated more favorably than unattractiveones throughout life—they are more popular, more sought after,more valued by lovers and friends, and so forth.Several studies have explored correlations between globalself-esteem and self-rated attractiveness, generally finding verystrong positive relationships. Harter (1993) described resultsfrom multiple studies indicating that the correlation was around.85. This is a remarkably strong connection, indicating thatpeople’s physical attractiveness accounted for more than 70%of the variance in their self-esteem. If this result is correct, itmeans that people’s self-esteem is mainly based on their appearance.But one cannot easily rule out the possibility that thecorrelation received an unfair boost by the general tendency tospeak well of oneself. People who score high on self-esteem byclaiming to be wonderful people in general may claim to bephysically attractive, and people who rate themselves relativelypoorly overall may derogate their appearance.Hence, it is important to obtain more objective measures ofphysical appearance, to match up with the subjective measuresof self-esteem. Diener, Wolsic, and Fujita (1995) obtained selfesteemscores from a broad sample of individuals and thenVOL. 4, NO. 1, MAY 2003 7

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemphotographed everyone. The photographs were shown to otherjudges, who rated them for attractiveness, thus providing anobjective measure of physical attractiveness unaffected by selfreport.The ratings of attractiveness based on full-length picturesshowed a trivial correlation of .06 with self-esteem—noteven a significant correlation. Head-and-shoulders photos faredonly slightly better, with a correlation of .14, which suggeststhat physical attractiveness accounted for less than 2% of selfesteem.Even this figure may be inflated, because it could bethat people with high self-esteem take particular care to presentthemselves in a favorable light, such as by wearing attractiveclothing and jewelry. When these props were removed to produceunadorned photos of the participants’ faces, the correlationof judge-rated attractiveness with self-esteem was .00. Inthat same investigation, however, self-reported physical attractivenesswas found to have a hefty (r .59) correlation withself-esteem. Thus, people’s ratings of their own good looks accountedfor 35% of the variance in their self-esteem, whereasobjective ratings of their looks had a negligible relationship totheir self-esteem. People with high self-esteem are gorgeous intheir own eyes, but objective observers do not see any difference.Similar findings were reported by Gabriel, Critelli, and Ee(1994), who obtained peer ratings of attractiveness from photographs.The correlations with self-esteem were negligible: r .01 for males and r .04 for females. Again, though, peoplewith high self-esteem rated themselves as more attractivethan those with low self-esteem.The discrepancy is sobering. What seems at first to be apowerful relationship between physical good looks and highself-esteem turns out to be nothing more than a pattern of consistencyin how favorably people rate themselves. Those whosay they are very good overall tend to say that they look good,too. When other people are brought in to judge unadorned photographs,however, people with high self-esteem do not emergeas any more attractive than people with low self-esteem.Inflated views of one’s own attributes are not confined tophysical attractiveness. Gabriel et al. (1994) also asked participantsto rate their own intelligence and then gave them anintelligence (IQ) test. People with high self-esteem rated themselvesas significantly more intelligent than people with lowself-esteem (r .35). But the results of the objective IQ testdid not justify these favorable claims, for there was no significantrelationship between self-esteem and IQ scores (r .07).The authors also reported that self-esteem was significantlycorrelated (r .38) with the discrepancy between self-rated intelligenceand objectively measured intelligence. This findingconfirms that people with high self-esteem exaggerate their intelligencemore than people with low self-esteem.Weight has long been associated with self-esteem, especiallyin modern Western cultures that glorify slender, younglookingbodies (especially for women). Hence, a lighter bodyweight should be associated with high self-esteem, whereasobesity should be linked to low self-esteem. Consistent withthat view, a meta-analysis (which statistically combines the resultsof multiple studies) by C.T. Miller and Downey (1999)found a significant correlation (r .24) between actual bodyweight and self-esteem. But the correlation of self-esteem withself-rated body weight was much stronger (r .72). Thus,people with high self-esteem are a little slimmer than others,but not nearly as much as they think.The broad implication of these examples is that self-reportsare likely to contain substantial biases that can easily yield misleadingempirical findings. People with high self-esteem claimto be successful, attractive, and wonderful in many respects.Objective evidence sometimes paints a much different picture,and many of the ostensible (self-reported) advantages claimedby people with high self-esteem are clearly disconfirmed.The systematic discrepancies between objective and selfreportedoutcomes led us to conclude that we should set up ourliterature search and review to emphasize objective measuresof outcomes whenever possible. This vastly reduced theamount of material we could use. But the material that remainedis far more reliable and convincing insofar as the resultsare based on objective measures.To be sure, objective measures are not always possible toobtain, and reliance on self-report is thus inevitable in somespheres. For example, in the case of happiness, there is almostno alternative to self-report. Even when it is necessary to useself-report, however, we urge researchers to emphasize themost objective, concrete, and verifiable data possible. It seemslikely, for example, that global self-ratings of intelligence aremore vulnerable to bias and inflation than self-reports of gradesin specific courses or scores on specific tests. Behavioral selfreports(e.g., “How many cigarettes did you smoke yesterday?”)should be more reliable than vague or “in general” ratingsof one’s own behavior (e.g., “How much do you smoke?Very much, some, not very much, or not at all?”).The bias in self-report may well be partly responsible forthe popularity of self-esteem among teachers, parents, therapists,and others who seek to intervene in people’s lives in apositive fashion. A rise in self-esteem may well cause a personto honestly believe that he or she is doing better in manyspheres, even if these beliefs are utterly false and stem from thepositive illusions that attend high self-esteem. If both teacherand student believe that boosting the student’s self-esteem hasled to improvements in academic performance, the entire enterpriseof boosting self-esteem is likely to be marked by a seductivefeeling of success and efficacy. The unfortunate fact thatacademic performance may not have changed at all by objectivemeasures might be overlooked in the context of the mutuallysustained social reality of the belief that performance hasimproved.Direction of CausalityA great many researchers have proceeded by administeringa self-esteem scale and some behavioral measures and thencomputing the correlation. Such an approach can establish8 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.links between self-esteem and other variables. Correlationscannot establish causality, however, and so the meaning of suchcorrelations remains highly ambiguous. As anyone who hastaken a research methodology course knows, a correlation betweenX and Y could mean that X causes Y, that Y causes X, orthat some other variable causes both. Thus, if self-esteem correlateswith good school performance, self-esteem might be eitherthe cause or the result of good performance, or bothoutcomes could derive from something else, such as a privilegedupbringing.The question of causality goes to the heart of the debateabout the self-esteem movement and interventions aimed atboosting self-esteem. If self-esteem causes an assortment ofpositive outcomes, then it may well be worth considerable effortand expense to instill high self-esteem into children, thementally ill, stigmatized populations, and other vulnerable orat-risk groups. In contrast, if the correlations mean simply thatself-esteem is an outcome of success and good behavior—which is, after all, at least as plausible as the hypothesis thatself-esteem is a cause of success—then there is little to begained by raising self-esteem.If self-esteem is indeed an outcome rather than a cause, thenit is even plausible that raising self-esteem could backfire andproduce undesirable effects. Suppose, for example, that workinghard in school leads to good grades, and good grades leadto high self-esteem. Assume also that high self-esteem feelsgood, and so the rise in self-esteem could operate as an importantreinforcer for the hard work that leads to academic success.If a school program intervenes directly to boost selfesteemregardless of academic performance, then students canenjoy the rewards of self-esteem without making the effort.One major incentive to work hard would thereby be eliminated,and students might reduce their effort, leading to poorer academicperformance. Such speculations indicate the need to attendto the causal processes involved in self-esteem and raisesubstantial questions about the impact of boosting self-esteem.The claim that correlational findings can shed no light oncausal processes is overstated. Two patterns of correlationalfindings are especially helpful for evaluating causality. First, ifA causes B, or B causes A, then A and B will show up as correlatedin most data sets. If A fails to show a significant correlationwith B, then it is fair to conclude that there is no causalrelationship in either direction (assuming that measurement erroror methodological shortcomings are not responsible for thenull correlation). Put another way, null correlations falsifycausal hypotheses.Second, causes generally precede consequences in time, andso correlations across time are often used to make causal inferences.If self-esteem on one occasion (Time 1) predicts schoolperformance at a later time (Time 2), whereas school performanceat Time 1 does not predict self-esteem at Time 2, thenone can reasonably conclude that self-esteem is the cause andschool performance is the outcome. Likewise, if self-esteem atTime 1 predicts school performance at Time 2 after differencesin school performance at Time 1 are controlled statistically (ineffect, self-esteem at Time 1 predicts change in school performancebetween Times 1 and 2), then causal inferences may bemade, albeit somewhat tentatively.Insisting that self-esteem must predict achievement at Time2 after controlling for achievement at Time 1 could obscuresome actual causal relationships, so it should be regarded as ahighly conservative way of testing the hypothesis. Suppose, forexample, that self-esteem does consistently contribute to achievement—andthat self-esteem is fairly stable. Achievement atTime 1 is thus a result of prior self-esteem (say, at Time 1). Ifone controls for achievement at Time 1 when computing thecorrelation between Time 1 self-esteem and Time 2 achievement,one may be throwing a very large baby out with the statisticalbathwater. What in reality was a replication of causalprocesses (Time 1 self-esteem causing Time 1 achievement,and Time 1 self-esteem causing Time 2 achievement) ends uplooking like no relationship at all.In short, care must be taken to avoid either overstating orunderstating the causal influence of self-esteem.METHOD OF SEARCHING THE LITERATUREThe research covered in the subsequent sections of this reviewwas located in the following manner. On several occasionsfrom January 2001 through October 2001, we searchedthe PsychINFO database and obtained a list of all articles containing“self esteem” in the abstract. Our searches yielded a totalof 15,059 publications. We also went to the “subject”category in PsychINFO and searched for “self esteem,” findinga total of 11,860 articles. Because these two lists overlappedsubstantially, we decided to rely mainly on one of them; wechose the former because it included more entries.The 15,000-article list was too long to work with, so we narrowedit down using a series of criteria. First, we did not retainwork that had not been published and whose sole source wasDissertation Abstracts International. This biased our results infavor of positive effects of self-esteem, insofar as null resultsare difficult to publish. Second, we excluded studies designedto show what causes self-esteem. In other words, we sought tounderstand the consequences of self-esteem rather than its determinants.Third, we generally eliminated secondary sourcesin order to obtain original data. Many book chapters, books,and other writings talk about self-esteem, but our goal was tosee what the original data indicate. We therefore included secondarysources only if they presented meta-analyses or otherauthoritative literature reviews.Three other sources of information beyond the PsychINFOlist were utilized. First, we have all been reading and writingabout self-esteem for many years, and we all felt it appropriateto contribute any additional work from our own knowledge.Second, sources cited in the articles from the PsychINFOsearch were also included. This was especially helpful for findingnull effects, because researchers who investigate self-esteemVOL. 4, NO. 1, MAY 2003 9

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemand fail to confirm predictions often cite previous work to notethat their own failure is not unique. Third, in August 2001, wesent out a request on the Listserve of the Society for Personalityand Social Psychology asking for any new or unpublishedarticles on objective outcomes of self-esteem. This requestelicited an additional five articles.Finally, we sought to narrow our list to include only workon outcomes with broad social relevance. We searched for specificrelationships between self-esteem and variables of interest,such as health, sexual behavior, financial status, grades,intelligence, job performance and satisfaction, and interpersonalrelations.We then read the remaining articles carefully. At this point,our overriding goal was to obtain studies that included objectiveevidence of the effects of self-esteem. As we have alreadynoted, some variables (such as happiness) can be assessed onlyvia self-report. Whenever possible, however, we wanted objectiveor third-party evidence. We also favored studies thatsought to establish causal relationships, either by longitudinaldesigns or by laboratory experimentation.Given our restrictive criteria, relatively few studies survivedfrom the initial set. Although many thousands of publicationsrefer to self-esteem, relatively few of them report studies thatused highly rigorous methods to examine the causal impact ofself-esteem on personal and social problems. These few generallyrepresent vast amounts of work and deserve careful attention.We summarize them individually rather than meta-analytically.SCHOOL PERFORMANCEThe self-esteem movement has been especially influential inAmerican schools, and part of the reason for this is the assumptionthat raising self-esteem will lead to improvements in children’sacademic performance. There are plausible reasons forthinking that high self-esteem will lead to good schoolwork.People with high self-esteem may set higher aspirations thanpeople with low self-esteem. They may be more willing to persistin the face of initial failure and less likely to succumb toparalyzing feelings of incompetence and self-doubt. Learning,by definition, involves acquiring information and skills that onedoes not initially have, and high self-esteem may help preventthe recognition of one’s initial incapability from producing asense that the cause is hopeless. High self-esteem may fosterthe confidence to tackle difficult problems and enable people toderive satisfaction from progress and success.Correlational FindingsMany studies have found that self-esteem is positively correlatedwith academic performance. In an early review, Wylie(1979) concluded that the correlation between self-esteem andstudents’ grade point averages was about .30. She added thatsimilar or slightly stronger relationships had been reported betweenself-esteem and scores on various achievement tests.Creativity, however, was not consistently related to any form ofself-regard.The most definitive compilation was Hansford and Hattie’s(1982) meta-analysis of 128 studies involving more than200,000 participants. These studies explored a variety of measuresof self-regard (mostly self-esteem) and a variety of objectiveperformance measures, most of which were achievementtests. The correlations reported varied widely, from .77 to.96, and averaged between .21 and .26 (depending onhow the average was computed, how studies were weighted,etc.). Hansford and Hattie concluded that overall there is a significantpositive relationship between self-esteem and academicperformance, with self-esteem accounting for between 4and 7% of the variance in academic performance.More recent studies have yielded similar conclusions. Usingstandard achievement tests, Davies and Brember (1999) foundsignificant though weak positive relationships between selfesteemand academic performance in a large (N 3,001) Britishsample. The correlations ranged from .10 to .13, and averaged.12. A somewhat stronger relationship was found by Bowles(1999), who showed that self-esteem correlated at .29 with students’most recent semester grades in mathematics and English.Kugle, Clements, and Powell (1983) found that scores ona reading achievement test correlated .18 with level of selfesteem.However, when these authors controlled for ethnicity, theeffect of self-esteem was no longer significant. Thus, these recentstudies also indicate that self-esteem goes with doing wellin school, although the relationship is weaker than one mighthave expected in a society that values doing well in school.As already noted, people with high self-esteem report theirintelligence to be high, although there is no relationship betweenself-esteem and scores on objective IQ tests (Gabriel etal., 1994). Simon and Simon (1975) found scores on self-esteemto be correlated significantly (r .33) with scores on academicachievement tests and also with IQ test scores (thus contradictingthe null result obtained by Gabriel et al.). In general,though, there is very little evidence that self-esteem correlateswith IQ or other academic abilities.Self-esteem also predicts performance in minority and atrisksamples. Ortiz and Volloff (1987) found significant correlationsbetween self-esteem and tests of IQ and school abilities,using a limited sample of Hispanic students in grades 3 to 6who had been nominated for testing as candidates for giftedclasses. Howerton, Enger, and Cobbs (1994) studied at-riskBlack male students, and found that self-esteem predictedgrades and school achievement. Although Howerton et al. usedobjective measures (school records) of achievement, a drawbackof the study is that the sample was very small (N 42).Different authors have drawn very different conclusionsfrom correlations between self-esteem and school performance.Among the most optimistic was that of Zimmerman,Copeland, Shope, and Dielman (1997): “Efforts either to preventand stabilize decreasing self-esteem or to build self-esteem10 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.may have vital effects on . . . outcomes for youth” (p. 137). Intheir own study, Zimmerman et al. measured performance byasking students for general ratings of their grades (e.g., “Areyour grades mostly A’s, mostly A’s and B’s . . . ?”), so their resultswere vulnerable to subjective bias. In addition, the correlationswith self-esteem were very weak. Their results do notseem to support their confidence that high self-esteem leads tobetter grades.Other authors have inferred that significant correlations betweenself-esteem and school performance support the oppositecausal conclusion, namely, that good work in school leadsto high self-esteem. Bowles (1999) specifically measured selfesteemafter the semester for which he obtained school grades,so the positive correlation he found seems most consistent withthe view that self-esteem is a result, not a cause, of doing wellin school. This conclusion was further supported by path analysis(a statistical technique for testing theories about complexchains of causes), which indicated that there was no directcausal path from self-esteem to achievement.Still other researchers have concluded that the correlationsbetween self-esteem and school performance, albeit significant,are so small as to be not worth pursuing. Rubin, Dorle,and Sandidge (1977) found that self-esteem was significantlycorrelated with all their measures of achievement, as well aswith teachers’ ratings of students’ behavior and performance.Yet statistical analysis showed that taking self-esteem into accountbarely improved the accuracy of predictions of achievementthat were based on socioeconomic status and intelligence(IQ) alone. Rubin et al. wrote, “While these increases were significant,their practical significance is negligible” (p. 506), andthey concluded that the links between self-esteem and academicperformance are based on “common underlying factorssuch as ability and background” (p. 503).These and other findings generally point to a positive butweak and ambiguous relationship between self-esteem andschool performance. Students with high self-esteem generallyhave done somewhat better in school and on school achievementtests than students with low self-esteem. The correlationalfindings do not indicate whether self-esteem is a cause ora result of school performance. They do, however, furnish onepossible explanation for the continuing belief that self-esteemmay be beneficial for school performance. Teachers, parents,and others may observe that high self-esteem and good schoolperformance go together and infer that self-esteem plays acausal role. Unfortunately, impressions—even when backed upby significant correlations—do not justify causal conclusions.We now turn to studies that have investigated whether there is acausal relationship between self-esteem and academic performance.Investigating CausalitySeveral studies have investigated the time course of the positiverelation between self-esteem and academic performance,as a way of establishing causal priority. Some of these studies havealso investigated whether third variables, such as socioeconomicstatus or intelligence (IQ), could be responsible for thecorrelations between self-esteem and academic performance.An early and still well respected study by Bachman andO’Malley (1977) used data from a nationwide longitudinalstudy that tracked more than 1,600 young men from 1966,when they were in 10th grade, up through 1974. All participantsin this study completed a modified version of the Rosenberg(1965) self-esteem scale at several points during thisperiod. Although Bachman and O’Malley found that self-esteemcorrelated with school performance, their more sophisticatedstatistical tests (i.e., path analyses) did not point to anycausal role for self-esteem. Instead, they concluded that sharedprior causes, including family background, ability, and earlyschool performance, affect self-esteem and later educational attainmentand were responsible for the correlation between thetwo. They also concluded that occupational success causedself-esteem to rise, whereas obtaining higher education had anegligible impact on self-esteem. Of all their findings, the onethat gives the most credence to the view that self-esteem is animportant cause of successful outcomes is that self-esteem inhigh school predicted eventual level of educational attainment(final degree earned), but in their path analysis the direct linkfrom high school self-esteem to later educational attainmentwas only .072. Its link to eventual occupational status was similar,at .061. These numbers indicate that the relationship is extremelyweak, if it exists at all. Neither link was significant.The findings of Bachman and O’Malley (1977) are importantfor several reasons. First, these researchers were perhaps the firstto conduct such a thorough and sophisticated study of the impactof self-esteem. Second, they plainly had hoped to find that selfesteemplayed a causal role, and they favored this hypothesis intheir initial exposition. Indeed, they proposed that high self-esteemfosters high aspirations and persistence, which lead ultimatelyto better academic performance and career success.Third, most subsequent work has generally confirmed their conclusionsthat self-esteem is a result rather than a cause, and thatany correlations between self-esteem and achievement are likelyto be due to third variables such as family background.Another milestone study was conducted by Maruyama, Rubin,and Kingsbury (1981). This research is an important complementto Bachman and O’Malley’s (1977) study because itfocused on a much younger age: Maruyama et al. followed a finalsample of more than 700 students from age 4 to age 15.Achievement was measured using academic achievement tests,including the Stanford Achievement Test and, later, the WideRange Achievement Test, which emphasizes spelling, vocabulary,and arithmetic. Like Bachman and O’Malley, Maruyamaet al. found that self-esteem and academic achievement are correlated,but concluded that there is no causal relationship betweenthose variables. Instead, they argued that ability (IQ) andsocial class are the underlying causal factors that affect the levelsof both self-esteem and academic achievement.VOL. 4, NO. 1, MAY 2003 11

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-EsteemA follow-up by Bachman and O’Malley (1986) confirmedtheir previous findings. In this follow-up, they sought to improvetheir methodological rigor in several ways. One of thesewas to confine analyses to White males in White-majorityschools. Although the exclusion of minorities is not consideredmethodologically desirable today, Bachman and O’Malley hadquite persuasive reasons for deciding that such inclusion mighthave weakened the chance for self-esteem to emerge as causallysignificant. Specifically, in their original sample, AfricanAmerican students scored lower than White students on achievementand ability tests but higher on self-esteem. Combining datafrom White and African American students could therefore obscurea positive relationship between self-esteem and achievement.The authors also acknowledged that the low achievementscores of African American students might indicate some lackof validity of those measures for such a sample. Any lack ofvalidity would make self-esteem look less effective than it actuallyis.Despite their effort to increase the potential for their studyto demonstrate effects of self-esteem, Bachman and O’Malley(1986) found that global self-esteem had a negligible relationshipto eventual educational attainment. Family background(socioeconomic status), ability (IQ), and early school gradespredicted eventual level of educational attainment, and selfesteemadded little to the accuracy of prediction. Self-esteemwas correlated with actual ability, although self-rated abilitywas consistently inflated: People thought they were smarterthan they actually were. In short, these findings support the roleof illusion in self-esteem, but they contradict the view that selfesteemcauses long-term educational success.An article titled “Is There a Causal Relation Between Self-Concept and Academic Achievement?” was published by Pottebaum,Keith, and Ehly (1986). Their answer was a ratherblunt negative: “The results suggest that there is no significantcausal relation between self-concept and academic achievement[in either direction], but rather that the observed relationis the result of one or more uncontrolled and unknown thirdvariables” (p. 142). A great deal of methodologically sophisticatedwork went into producing that conclusion. Pottebaum etal. used a sophisticated research design, testing a very largesample of high school students (more than 23,000) in the 10thgrade and again in the 12th grade. Self-esteem in 10th gradepredicted academic achievement in 12th grade quite weakly (r .11). Conversely, academic achievement in 10th grade predictedself-esteem in 12th grade only trivially better (r .12).The authors noted that a reciprocal causal relationship betweenself-esteem and academic performance could produce the patternof findings they obtained—but only if the two variablescause each other with about the same amount of power, whichseemed a priori implausible. Hence, Pottebaum et al. concludedit is more likely that self-esteem and academic performanceare both the result of a third variable (or set ofvariables). This conclusion is clearly consistent with whatBachman and O’Malley (1977) and Maruyama et al. (1981)found, despite different methods and somewhat different numericalresults.A slightly different conclusion was reached by Rosenberg,Schooler, and Schoenbach (1989). They analyzed data from theYouth in Transition longitudinal study that was also the basisfor Bachman and O’Malley’s (1977) work. They used the 10thand12th-grade data for nearly 1,900 boys, including measurementsof achievement that relied on having the students reporttheir grade point average in school. This sort of measure is inthe middle of the span of methodological rigor that we haveoutlined: It is not fully objective, but it is somewhat specificand verifiable. A slight tendency of people with high selfesteemto furnish self-flattering reports is to be expected, althoughthe scope for such inflation may be more limited than ifrespondents are asked a general self-rating question, such as“How good are you at school?”Rosenberg et al. (1989) did find significantly positive, althoughweak, correlations between self-esteem and self-reported grades,r .24 in 10th grade and r .25 in 12th grade. Of greater interestwere the findings across time. These supported the conclusionthat self-esteem is the result of grades, rather than thecause. There was a modest causal relationship (.15) leadingfrom grades to self-esteem, but the causal relationship leadingfrom self-esteem to grades was only .08, which was not significantlydifferent from zero. In other words, there was no solidevidence that self-esteem had any effect on grades, despite thefact that even a weak relationship would likely have been significantbecause Rosenberg et al. tested so many people.The view that self-esteem is an outcome rather than a causeof good school performance was further supported in anotherlarge and sophisticated study, by Skaalvik and Hagtvet (1990).Their sample consisted of 600 Norwegian schoolchildren intwo cohorts, one in third grade (about age 9) and the other insixth grade at the start of the study. A second set of data wasobtained a year and a half later. Achievement was measured byteachers’ ratings, which furnish a good measure although erroneousperceptions by the teacher could affect a student’s selfesteem(e.g., if the teacher treats the student as a genius or dullard).Skaalvik and Hagtvet found evidence that doing well inschool one year led to higher self-esteem the next year,whereas high self-esteem did not lead to performing well inschool. In fact, high global self-esteem in grade 6 predictedlower academic achievement in grade 7.Skaalvik and Hagtvet (1990) also measured students’ selfconceptof their academic ability. Although our focus is not ondomain-specific measures of self-esteem in this review, thefindings are of interest. These researchers concluded that selfconceptof ability mediates the relation between academic performanceand global self-esteem. Specifically, doing well inschool leads to thinking of oneself as good at schoolwork,which in turn can boost global self-esteem. There was alsosome evidence for a causal influence of academic self-concepton school performance, which is thus one finding that suggeststhinking well of oneself can lead to better schoolwork (al-12 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.though this relationship involves thinking of oneself as good atschoolwork rather than as good overall). This causal influencewas found in some analyses but not others and was not strong.Still, the fact that it was demonstrated at all helps dispel worriesthat some kind of methodological or measurement problemrenders it impossible to verify causal effects of self-concept.This in turn lends further credence to the conclusion that globalself-esteem is not a cause of school performance.InterventionsWe found relatively little evidence on how self-esteem programsor other interventions affect self-esteem. Such interventionsare practiced in many schools and other places, but it iscommon for them to target not only self-esteem but also studyskills, citizenship, conflict reduction, and other variables. Obviously,if a program that attempts to boost self-esteem and improvestudy skills ends up producing an improvement ingrades, it is hardly safe to conclude that self-esteem is responsiblefor the improvement. Furthermore, given that the studiesinvestigating causality have not demonstrated that self-esteemhas an impact on academic achievement, it seems likely thatany attempt to collect solid data on the impact of boosting selfesteemwould end up with null results, and the evaluatorswould not be anxious to publish their results even if they could.An impressive review of research on such programs waspublished by Scheirer and Kraut (1979). The title, “IncreasedEducational Achievement via Self-Concept Change,” soundspromising with respect to the benefits of self-esteem, but thefindings were not. Scheirer and Kraut covered evidence fromboth published and unpublished evaluations of school-basedinterventions and programs, including Head Start, the EarlyTraining Project, and Upward Bound. By and large, the evaluationsof these studies “generally failed to find an associationbetween self-concept change and academic achievement” (Scheirer& Kraut, 1979, p. 140). When there was an association, ittended to be temporary (i.e., it was not maintained beyond theend of the program), or it indicated that self-esteem was the resultof academic achievement rather than the cause. Programsthat targeted factors other than self-esteem (such as by encouragingparents to become involved in their children’s schoolwork)seemed to get better results.Scheirer and Kraut (1979) carefully considered a variety offactors that could have led to the general pattern of null results,including poor measurement, methodological problems, andfailure to implement interventions properly, among others. Yetas far as they could tell, these factors were not sufficient to explainthe broad pattern of results. For example, they noted thatsome interventions that did not target self-esteem had producedmeasurable gains in academic performance. Hence, they concludedthat the most likely explanation of the disappointing resultswas that the basic theoretical hypothesis—namely, thatimproving self-esteem will lead to better academic performance—waswrong. They said that self-esteem may be an outcomeof academic achievement, but it does not appear to be acause or a mediating variable.A nicely controlled field experiment by Forsyth and Kerr(1999) provided converging evidence using an adult (collegestudent) sample. This investigation was conducted in connectionwith a regular course. Students who received a C, D, or Fon the first examination were targeted to receive weekly e-mailmessages from the professor. Each message contained a reviewquestion pertaining to that week’s assignment. In the controlcondition, the review question was all that was included in thee-mail. Students who were randomly assigned to two othergroups, however, received either a message aimed at boostingtheir sense of personal control and responsibility for their ownperformance or a message aimed at boosting and maintainingtheir positive sense of self-worth. These manipulations had noeffect on the C students, but the D and F students who receivedthe self-esteem boost performed significantly worse on subsequenttests than the D and F students in the other conditions.Forsyth and Kerr’s (1999) study is of interest because it involvedfull random assignment and because it found significantdifferences as a function of treatments aimed at self-esteem.Unfortunately for the self-esteem movement, the findings suggestthat the intervention aimed at boosting self-esteem wascounterproductive in its impact on academic performance. Wehave already noted some evidence linking high self-esteem tosubsequently poorer performance (e.g., Skaalvik & Hagtvet,1990), although the preponderance of findings suggests thatself-esteem is positively linked to academic achievement. Still,Forsyth and Kerr used an intervention rather than simple measurementof current self-esteem, and they aimed their interventionspecifically at low-performing students. Their results areconsistent with the view that self-esteem is an important reinforcerfor good academic performance and that supplying thereward indiscriminately (i.e., not linking it to good performance)may deflate its reward value. In plainer terms, studentsmay ordinarily work hard in order to be permitted to feel goodabout themselves, and an intervention that encourages them tofeel good about themselves regardless of work may remove thereason to work hard—resulting in poorer performance.ConclusionThe impact of self-esteem on school performance has beenstudied more carefully and thoroughly than any other outcomewe discuss in this review. We were able to find studies that usedlongitudinal designs and other methods that can address causation,used objective methods rather than relying on self-report,and attempted to untangle self-esteem from other variables.The results do not support the view that self-esteem has astrong causal effect on school achievement. Indeed, most of theevidence suggests that self-esteem has no impact on subsequentacademic achievement. The few studies suggesting anypositive causal impact of self-esteem generally found only tinyVOL. 4, NO. 1, MAY 2003 13

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemeffects. Some findings even point (again weakly) in the oppositedirection, suggesting that high or artificially boosted selfesteemmay detract from subsequent performance.There were in fact some reasons to hope that self-esteemwould be more potent. On a theoretical basis, self-esteemseemed likely to enhance academic strivings and persistence.Early empirical findings may also have encouraged the beliefthat self-esteem is helpful, because simple correlations betweenself-esteem and academic performance have often beenpositive and significant (as noted by Wylie, 1979). Unfortunately,those correlations appear to be due to processes otherthan self-esteem causing good performance. First, good performancein school may sometimes lead to higher self-esteem, insteadof the reverse. (Even that tendency is disappointinglyweak, however.) Second, self-esteem overlaps with other variables,and when these are controlled in the statistical analyses,the proportion of variance in performance accounted for byself-esteem dwindles rapidly.JOB AND TASK PERFORMANCEApplied to the world of work, the self-esteem hypothesissuggests that people who feel better about themselves performbetter. Early reviews of studies on the relationship between selfesteemand work performance noted the high variability of thereported size and nature of this relationship (Brockner, 1983;Tharenou, 1979). Judge and Bono (2001) surveyed the resultsof 40 studies with more than 5,000 participants and found thatmost of them showed weak positive relationships. In some studies,self-esteem was measured with global indices, such as theRosenberg scale, whereas in others, it was measured more specificallyas organization-based self-esteem. Similarly, there weredifferences in the way performance was measured. Some studiesemployed self-perception measures, whereas others used supervisors’ratings or objective measures. Whether the associationbetween self-esteem and performance was moderated by suchdifferences in measurement was not reported.Ultimately, the results regarding the link between self-esteemand job performance seem to echo what has been found withschool performance: The simple search for correlations yields amodest positive relationship, such that high self-esteem goeswith better performance. However, the reported link betweenself-esteem and job performance is highly variable, in contrastto the more consistent findings regarding school performance. Alikely explanation is that the relevance of self-esteem to workperformance varies substantially because the demands and rewardsof different occupations are highly variable.The most important point, however, is that correlationalfindings do not permit causal inferences. Self-esteem mightimprove job performance. Then again, occupational successmight well boost self-esteem, whereas failure at one’s job maydeflate it. As in the case of school performance, time-laggedlongitudinal studies are useful for determining whether selfesteemis a result or a cause of good performance, but thesehave not yet been done.A quite different approach to studying how self-esteem mayinfluence performance is to investigate the relationship in thelaboratory. Such studies typically measure self-esteem as a preexistingtrait and then investigate task performance in the laboratory.In most cases, researchers manipulate other variablessuch as initial failure or size of reward to investigate how theyinteract with self-esteem.Performance QualityRelatively few studies on the effect of self-esteem on performancequality have been published, and this leads us to suspectthat the results are typically nonsignificant. We can point to evidencefrom our own laboratories. Wallace and Baumeister(2002) had students solve arithmetic problems under varyinglevels of challenge and performance pressure. There was noeffect of self-esteem on any of the performance measures.Baumeister, Heatherton, and Tice (1993) examined performanceat a video game that was presented as a test of nonverbal intelligence.Across multiple studies, there was essentially no effectof self-esteem on performance. The control conditions (whichindicate whether there is a link between self-esteem and performancein the absence of special interventions) typically showedno difference in performance as a function of self-esteem,whereas ego threat in the form of critical feedback on a priortest caused people with high self-esteem to perform worse thanthose with low self-esteem.In contrast, J.D. Campbell and Fairey (1985) found thatpeople with high self-esteem performed better than those withlow self-esteem on an anagram test when participants were firstasked to write explanations for a hypothetical failure (therewere no self-esteem differences when participants first wroteexplanations for a hypothetical success or in a control condition).Thus, when the prospect of a possible failure is salient,higher levels of self-esteem may benefit task performance. Ingeneral, though, self-esteem seems to have little or no direct relationshipto task performance.Persistence at TasksConventional wisdom claims that persistence is a vital aspectof performance (at both school and work). The research onpersistence has generally been consistent with the notion that acore characteristic of people with high self-esteem is their tendencyto persist in the face of failure (e.g., McFarlin, Baumeister,& Blascovich, 1984; Perez, 1973; Shrauger & Sorman,1977, Study 1).The conventional wisdom notwithstanding, blind persistenceis not an ideal self-regulatory strategy; there is also valuein knowing when to quit. Some studies have not found a correlationbetween greater persistence and high self-esteem, butthese seeming failures to replicate suggest that individuals with14 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.high self-esteem may make better use of cues telling themwhether it is prudent to persist or withdraw. For example, Janoff-Bulmanand Brickman (1982) and McFarlin (1985) toldsome participants that some of the tasks that they would workon did not have solutions. This information caused high selfesteemparticipants to persist less (compared with the controlparticipants who did not receive this information), but did notaffect the persistence of low self-esteem participants. In a similarvein, Sandelands, Brockner, and Glynn (1988) advisedsome participants that persistence was a wise strategy and othersthat it was a less prudent strategy. People with high selfesteempersisted more in the former condition than in the latter,whereas those with low self-esteem were relatively unaffectedby the manipulation.Thus, when people are given advice or cues that persistencemay not be a good strategy, those with high self-esteem persistless than those with low self-esteem. Results were rather different,however, when McFarlin et al. (1984) directly told participantswhat to do when they encountered a difficult problem(quit or persist). In this case, low self-esteem participants wereresponsive to the instructions, whereas high self-esteem participantswere not. Taken together, these studies suggest that peoplewith high self-esteem make better use of situational cues indeciding for themselves the appropriate course of action, butthat people with low self-esteem may be more responsive to directionsthat simply tell them what to do.Although these experiments are interesting, in the realworld individuals are rarely (if ever) provided with this kind ofinformation. Indeed, persistence is a ubiquitous phenomenonprecisely because people are not given such direct advice. Theexperiments also differ from natural conditions in that experimentalparticipants experience only a single instance of failure,and there are no opportunities to pursue alternative goals ortasks (Bandura, 1989). Under these conditions, persistence atthe failed task seems to be a reasonable strategy. In contrast, ifone has experienced multiple failures on a task and, as is typicallythe case outside the laboratory, there are opportunities topursue alternative goals, it seems more adaptive to abandon theoriginal task in favor of an alternative.Di Paula and Campbell (2002) recently examined the relationshipbetween level of self-esteem and knowing when toquit. First, they conducted a laboratory experiment in whichthey manipulated the degree of failure (a single, initial failureor three instances of failure) and the availability of an alternativegoal. Compared with low self-esteem participants, highself-esteem participants persisted more after a single failure,but less after repeated failure when an alternative was available.In a second correlational study, participants enumerated10 specific goals at the beginning of an academic year and atthe end of the year indicated, for each goal, their perceivedprogress and behavioral persistence, as well as how much theyhad ruminated about the goal. Self-esteem was positively correlatedwith achieving more goals, more satisfaction withprogress toward goals, more behavioral pursuit of goals, andless rumination. More important, compared with participantswith low self-esteem, those with high self-esteem showedhigher correlations between their ratings of perceived progressand behavioral persistence across the 10 goals.Taken together, this research supports the notion that highself-esteem individuals utilize more adaptive self-regulatorystrategies than low self-esteem individuals; these strategiesmay, in fact, be a factor in their higher levels of reported happinessand satisfaction (reviewed in a later section). The fact thathigh self-esteem people have more functional responses to failuremay also convey slight advantages in performance, andthus contribute to the (weak) associations between self-esteemand performance.ConclusionHigh self-esteem has value in causing people to persistlonger in the face of failure, at least when there is no viable alternativetask or goal to pursue and when no cues suggest thatpersistence is not a good strategy. When there is an alternativegoal or information that persistence may be a poor strategy,high self-esteem is also associated with knowing when to quit.Thus, in performance contexts, high self-esteem people appearto use better self-regulation strategies than low self-esteempeople.It is difficult to draw firm causal conclusions about selfesteemand other aspects of job and task performance. Overall,there are weak positive correlations between job performanceand self-esteem, but these may be due in whole or in part eitherto successful performance improving self-esteem or to selfreportbiases. There is no strong evidence indicating specificallythat high self-esteem leads to improved performance on thejob. The links between self-esteem and job performance deservemore systematic and careful study, using objective measuresand longitudinal designs. However, if high self-esteemconsistently produced better performance in laboratory tasks,this would be well established by now, and the lack of such evidencesuggests to us that self-esteem has little impact on taskperformance.INTERPERSONAL RELATIONSAnother major sphere of life activity is interpersonal relations.The apparent failure of self-esteem to contribute directlyto improvements in academic and job performance would easilybe offset if self-esteem helped people to get along betterwith others.It is certainly plausible that high self-esteem would improveinterpersonal relations. People with high self-esteem claim tobe more popular than people with low self-esteem (Battistich,Solomon, & Delucchi, 1993), although as already noted, selfreportsare not reliable guides to objective realities. Highself-esteem might well make a person more likable insofar aspeople may prefer to interact with confident, enterprising indi-VOL. 4, NO. 1, MAY 2003 15

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemviduals and to avoid interacting with people who suffer fromself-doubts and insecurities. The reverse causal relationship isalso quite plausible, of course, and indeed it forms the centerpieceof the sociometer theory of self-esteem put forward byLeary and his colleagues (Leary, Tambor, Terdal, & Downs,1995; see also Leary & Baumeister, 2000; Leary & Downs,1995). According to this view, self-esteem operates as an internalmeasure of one’s interpersonal appeal and success. Popularitywould therefore cause self-esteem to rise, whereas socialrejection would cause it to drop. It is also plausible that highself-esteem elicits dislike, insofar as people reject conceited orarrogant individuals.Friendship Patterns and Interpersonal InteractionsAs already noted, people with high self-esteem claim to beespecially popular (Battistich et al., 1993). They also rate theirfriendships as being of higher quality compared with peoplewith low self-esteem (Keefe & Berndt, 1996). People low inself-esteem report more negative, aversive social interactions,more stressful life events, and less social support than peoplehigh in self-esteem (Lakey, Tardiff, & Drew, 1994). In theworkplace, people with high self-esteem claim to get along betterwith their co-workers (although not with their supervisors),compared with people with low self-esteem (Frone, 2000).Thus, self-reports indicate that the social lives of people withhigh self-esteem are far better, richer, and more satisfying thanthe lives of people with low self-esteem. But are these claimsjustified?In a large sociometric study by Bishop and Inderbitzen(1995), 542 ninth-grade students nominated their most-likedand least-liked peers. On the basis of these votes, each personwas put into one of five categories along a span of degree ofpopularity (i.e., popular, average, controversial, neglected, rejected).Perhaps surprisingly, these categories showed no correlationwith self-esteem whatsoever. This finding contradictsboth the view that high self-esteem is socially appealing and thetheory that self-esteem results from peer approval. One findingthat offered minimal support for the general self-esteem hypothesiswas that people who had no reciprocal friendships at all exhibitedlower self-esteem than people who had at least one.Beyond one reciprocal friendship, however, self-esteem appearedto be irrelevant to friendship status and popularity.Glendinning and Inglis (1999) categorized students intofour distinct social groups on the basis of their self-reported attitudesabout peer relations and their own popularity and thenrelated group membership to the students’ self-esteem levels.The social groups were called “peer oriented” (i.e., high status,popular with peers), “conventional” (i.e., middle status, acceptingof authority), “socially isolated” (i.e., hard to make andkeep friends), and “disaffected” (i.e., troublesome, disrespectfulof authority). Results showed that adolescents with lowself-esteem were relatively more common in the socially isolatedand disaffected categories, whereas those with high selfesteemwere more often found in the conventional and peerorientedcategories. Unfortunately, because these findings werebased on students’ self-rated popularity rather than peer ratings,they may reflect the tendency of people with high selfesteemto rate themselves favorably. Without objective measures,it is hard to know whether some students were actuallymore popular or merely believed themselves to be so.A somewhat better approach was taken by Dolcini andAdler (1994), who asked a subsample of eighth graders to indicatewhich peer group, or “crowd,” every other student in theeighth grade belonged to, and then obtained self-esteem scoresfrom all the students directly. Global self-esteem did not differamong the different crowds, although there were some effectsfor domain-specific self-esteem. For example, the academicallyoriented “smart” crowd scored higher than the jocks on academicself-esteem, and the athletic crowd scored higher thanthe bookworms on athletic self-esteem.Adams, Ryan, Ketsetzis, and Keating (2000) used teachers’ratings as the basis for evaluating the interpersonal behavior andsocial desirability of 300 schoolchildren in grades 4 and 7 inOntario, Canada. Two main types of desirable behavior were thefocus of the ratings: peer sociability, which included popularityand good relationships, and desirable classroom behavior,which included following rules and behaving properly in class.Neither popularity nor classroom behavior varied as a functionof self-esteem. Several background factors and child characteristicsdid successfully predict these outcomes, though (e.g., highscores on assertiveness were correlated with higher popularitybut also higher levels of rule breaking in class). Thus, the nullfindings for self-esteem do not appear to reflect a general failureof measurement or a high amount of random fluctuation. Selfesteemdid show some correlations with some of the backgroundvariables, such as family cohesion, but even thoughthese background variables predicted popularity and classroombehavior, self-esteem failed both as a possible mediator and asan independent predictor of the outcomes. The authors concludedglumly that their findings “call into question whetherself-esteem is a potential causal agent or even a correlate of interpersonalrelations or classroom citizenship” (p. 247).A smattering of methodologically sound studies with adultcollege students has examined the social interaction patterns ofpeople high in self-esteem. An important study by Buhrmester,Furman, Wittenberg, and Reis (1988) examined a variety of interpersonalbehaviors. All the students were rated both by themselvesand by their roommates, so the method offered a valuableopportunity to compare subjective and external perspectives. Asusual, people high in self-esteem rated themselves as significantlybetter than other people on all interpersonal skills: Theyclaimed to be substantially better at initiating relationships (r .63), better at disclosing things about themselves (r .41), betterat asserting themselves in connection with objectionable behaviorsby others (r .40), better at providing emotionalsupport to others (r .30), and significantly better even at managinginterpersonal conflicts (r .20).16 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.The roommates’ ratings told a very different story, however.For four of the five interpersonal skills, the correlation betweenself-rated self-esteem and roommate-rated skill fell short ofsignificance, ranging from .01 for conflict management to .15for assertion in the face of objectionable behaviors. In the caseof initiating new social contacts and friendships, the link toself-esteem (r .38) remained significant but was substantiallylower than the link between self-esteem and self-ratedskill at making new friends. A troublesome aspect of thisstudy’s findings was that roommate-rated self-esteem correlatedbetter than self-rated self-esteem with roommate-ratedsocial skills, and the correlations between roommate-rated selfesteemand roommate-rated interpersonal skills were quitesimilar to the correlations between self-rated self-esteem andself-rated interpersonal skills. Self-esteem almost certainly hasto be more accurate when reported by the self than when estimatedby others, and so these findings raise the worrisome possibilitythat the correlations were due to some kind of responsebias that resulted in answers to questions about self-esteem beingsimilar to answers to questions about interpersonal skills.The findings of Buhrmester et al. (1988) thus suggest thatpeople high in self-esteem regard themselves as having an impressiverange of superior social skills, but that these skills arenot readily discerned or confirmed by others. By and large,roommates did not find their high self-esteem roommates to beany more socially skilled than their low self-esteem roommates.The only significant exception involved skill in initiatingnew relationships and interactions. This does seem to beone sphere in which confidence could come into play: Peoplewho think that they are desirable and attractive might well bewilling to strike up conversations with strangers, possibly becausethey do not expect rejection. In contrast, people with lowself-esteem might shy away from initiating such contacts.Interpersonal interactions were studied in the laboratory byBrockner and Lloyd (1986). All subjects first rated their selfesteemand then had a 10-min “get acquainted” conversationwith a member of the opposite sex. Care was taken to ensurethat conversational partners did not know each other prior tothe session. Afterward, they answered questions about their impressionsof each other. In addition to these peer ratings, selfratingsof likability were obtained. People with high self-esteemestimated that their partners had liked them much morethan people with low self-esteem, but partners’ actual ratingsdid not show even a hint of this correlation. It is also noteworthythat people with low rather than high self-esteem appear tohave been responsible for the discrepancy between self-ratingsand partner ratings: People low in self-esteem estimated thattheir partners liked them much less than the partners actuallyindicated, whereas people high in self-esteem estimated theirpartners’ ratings quite accurately. These results dovetail withthose of Buhrmester et al. (1988): High self-esteem makes forthinking oneself popular and likable, but in reality people withhigh and low self-esteem are equally likable.Similar findings were reported by J.D. Campbell and Fehr(1990). After a 15-min get-acquainted conversation, people estimatedhow their partner and an observer would evaluate themand rated how favorably they evaluated their partner. Ratingswere also obtained from uninvolved observers. People withhigh self-esteem estimated that their partners and observerswould rate them more positively than people with low selfesteem,but there was no difference in the actual ratings bypartners or by observers. As in the study by Brockner andLloyd (1986), people high in self-esteem estimated quite accuratelyhow much their partners liked them (low self-esteempeople underestimated how much their partners liked them).However, because the observers rated everyone more harshlythan did interaction partners, low self-esteem people weremore accurate in their estimates of the observers’ ratings.A more complex laboratory investigation of self-esteem andinterpersonal liking was recently published by Heatherton andVohs (2000). They reasoned that the null findings of Brocknerand Lloyd (1986) might have been due in part to the relativelyneutral context, so they introduced a manipulation of egothreat. Specifically, prior to the conversational interaction, halfof the participants took a difficult intellectual test and receivedesteem-threatening feedback that they had performed far worsethan their peers. The other participants took a neutral version ofthe same test, which was labeled as a pilot task, and receivedno feedback. Then all participants had a structured, 20-minconversation with a naive participant involving topics of low,medium, and high intimacy. Finally, the participants rated howmuch they liked each other.Heatherton and Vohs’s (2000) results showed that self-esteeminteracted with ego threat. In the control condition, in whichthere was no blow to anyone’s pride or sense of competence,people were liked equally regardless of their level of self-esteem.In that respect, the results replicated the findings of Brocknerand Lloyd (1986) and J.D. Campbell and Fehr (1990). In theego-threat condition, however, differences did emerge. Peoplewith high self-esteem were given lower ratings than peoplewith low self-esteem. These findings directly contradict theview that high self-esteem leads to better interpersonal liking.(They also contradict the so-called buffer hypothesis, whichholds that people with high self-esteem cope better with stressand adversity. We discuss this hypothesis further in the sectionon Coping and Depression.) Additional evidence suggestedthat the decreased liking for people with high self-esteem in theego-threat condition compared with the control condition wasdue to their being perceived as antagonistic. In contrast, andperhaps counterintuitively, low self-esteem people were likedsignificantly more in the ego-threat condition than in the controlcondition.These findings were corroborated by Vohs and Heatherton(in press) in a naturalistic study of 372 college men and womenwho were rated by their peers. Participants’ global self-esteemwas assessed during the spring of their senior year in highschool and then again midway through their 1st year at an IvyLeague college. Those whose self-esteem dropped substan-VOL. 4, NO. 1, MAY 2003 17

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemtially (i.e., by 1 SD), presumably because the move to collegeshook their positive sense of self, were placed in the ego-threatgroup, and compared with students whose self-esteem did notdrop. This study supported Heatherton and Vohs’s (2000) previousstudies in showing that high self-esteem men and womenwho were threatened were liked least by their peers. Onceagain, being seen as antagonistic mediated the effect.Vohs and Heatherton (2001) looked into possible mechanismsunderlying these results. They found that high and lowself-esteem people think about themselves differently afterthreat, such that high self-esteem people become more independentand less interdependent (i.e., they emphasize self-relianceand believe that they really do not need other people),whereas the opposite is true for low self-esteem people (i.e.,they emphasize interpersonal relations, even at the expense ofthe self). In the absence of threat, however, high and low selfesteempeople show no differences in concern for the self versusinterpersonal relations. Further, Vohs and Heatherton foundthat it is seeing the self as separate versus connected to othersthat drives differences in likability and personality perceptionsbetween high and low self-esteem people who have experiencedan ego threat.Converging evidence was recently provided by Bonanno,Field, Kovacevic, and Kaltman (2002). They found that untrainedobservers disliked people prone to self-enhancement.Among the same group of participants, the size of the tendencytoward self-enhancement was associated with the self-reportedability to cope with stress and trauma (see the section on Copingand Depression). Put another way, holding an inflated viewof themselves helped people feel good and recover fromtrauma, but also tended to alienate others.Other evidence, however, suggests that people with highself-esteem are liked better if certain conditions are met. Joineret al. (1992) conducted a 5-week study of same-gender roommates,measuring roommates’ liking using a modified versionof the Rosenberg scale. Thus, if roommates rated people asthey rated themselves, people low in self-esteem would appearto be disliked, and people with high self-esteem would appearto be liked. This procedure therefore made it extra likely thatpeople with high self-esteem would emerge as well liked. Evenso, people with low self-esteem were liked less than othersonly if they were male, depressed, and prone to reassuranceseeking (an annoying pattern of interpersonal tactics designedto elicit positive, supportive comments from others). The resultsfor females, nondepressed individuals, and other participantsdid not show greater dislike or disparagement of peoplewith low self-esteem than people with high self-esteem.Romantic and Intimate RelationshipsThe role of self-esteem in romantic relationships has receivedfairly little attention. In particular, little is known aboutwhether self-esteem predicts the durability of romantic relationships.One study with a very small sample (N 30) foundthat couples with low self-esteem were more likely than coupleswith high self-esteem to break up over a 1-month period(S.S. Hendrick, Hendrick, & Adler, 1988). Data on love stylesand self-esteem support this finding, showing that low selfesteemis related to feelings of manic love, which is characterizedby extreme feelings of both joy and anguish over the loveobject (W.K. Campbell, Foster, & Finkel, 2002). High selfesteemis related to passionate, erotic love, which is marked bythe escalation of erotic feelings for the love object. These findingsare consistent with other studies showing that, comparedwith people with high self-esteem, those with low self-esteemexperience more instances of unrequited love (Dion & Dion,1975) and more intense feelings of love for others (C. Hendrick& Hendrick, 1986).Several findings indicate that relationship behavior differsas a function of self-esteem. Murray, Rose, Bellavia, Holmes,and Kusche (2002) found that people low in self-esteem engagein a variety of potentially destructive behaviors. They tendto distrust their partners’ expressions of love and support, andso they act as though they are constantly expecting their partnersto reject and abandon them. Thus far, however, these patternshave not translated into any evidence that the relationships areactually more likely to dissolve.Thus, despite the relationship problems caused by low selfesteem,relationships are no more likely to break up if a partnerhas low self-esteem than if a partner has high self-esteem. Possiblythe reason for this is that high self-esteem leads to relationshipproblems, too. Rusbult, Morrow, and Johnson (1987)examined four types of responses to problems within close relationships,and found that self-esteem produced the largest differencein the active-destructive (“exit”) category of responses.People with high self-esteem were significantly more likelythan others to respond to problems and conflicts by deciding toleave the relationship, seeking other partners, and engaging inother behaviors that would actively contribute to the deteriorationof the relationship. These results were based on responsesto hypothetical scenarios, which share many of the drawbacksof self-report measures. However, as the authors noted, it seemsunlikely that their findings can be attributed to a simple responsebias because people with high self-esteem were admittingto more undesirable, rather than desirable, behaviors.Shackelford (2001) found that self-esteem was intertwinedwith a variety of patterns in marriage, although he did not provideevidence as to whether high self-esteem affects the durabilityof marriages. Spouses showed similar levels of selfesteem,with global self-esteem of spouses correlating at .23and physical self-esteem (including self-rated attractiveness)correlating fairly strongly at .53. Significantly, Shackelford regardedself-esteem as an outcome rather than a cause of maritalinteractions, although his data were correlational. Wives’ fidelitywas the strongest predictor of husbands’ self-esteem. Thismight indicate that men with high self-esteem cause their wivesto remain faithful, or—as Shackelford speculated—that cuckoldedhusbands experience a loss of self-esteem.18 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.Most important, women complained more about husbandswith low than with high self-esteem. Low self-esteem menwere derided by their wives as jealous, possessive, inconsiderate,moody, prone to abuse alcohol, and emotionally constricted.Again, the direction of causality is difficult todetermine. Possibly, husbands’ low self-esteem elicits negativeperceptions among wives. Conversely, being disrespected ordespised by his wife may lower a man’s self-esteem. Yet anotherpossibility is that having a variety of bad traits leads bothto low self-esteem and to being disrespected by one’s wife.Meanwhile, the self-esteem of wives was unrelated to theirhusbands’ complaints about them, except that husbands whocriticized or insulted their wives’ appearance were generallymarried to wives with low self-esteem, and indeed Shackelford(2001) found that this was the most consistent predictor of lowself-esteem among wives.Group Behavior and LeadershipBehavior in groups is an important complement to behaviorin intimate dyadic relationships. We found relatively little workthat shows objective behavioral effects of self-esteem in groupsettings. One notable exception is an impressive longitudinalstudy by LePine and Van Dyne (1998). These researchers compileda broad sample of work groups, consisting in total ofnearly 600 employees in 21 different organizations. They followedthe groups for 6 months and obtained peer ratings of thebehavior of each participant. The main variable of interest was“voice behavior,” defined as the willingness to speak within thegroup, criticizing its operation or proposing alternative routes ofaction. Voice behavior is vital to an organization’s long-termsuccess, because innovation depends on it, as do the group’schances for stopping itself from proceeding down a destructivepath (as in “groupthink”; Janis, 1982). Because speaking up andcriticizing the group require initiative and a willingness to deviatefrom the group’s apparent consensus, the authors predictedthat high self-esteem would lead to greater voice behavior.This prediction was confirmed, although the effect was onlymarginally significant. Across the 6-month period, LePine andVan Dyne (1998) found that high self-esteem led to more voicebehavior, though it accounted for only 1% of the differencesbetween people on that dimension. Moreover, it was indeedpeople high in self-esteem who were most willing to speak out.An additional 3% of the variance in voice behavior was accountedfor by interactions of self-esteem with other variables,especially group size. As one might expect, as groups becamelarger, fewer people were willing to speak out against the consensusor criticize the group, but this pattern of declining initiativewas less pronounced for people high in self-esteem thanfor those low in self-esteem. Put another way, people with highself-esteem were less intimidated by the prospect of speakingout in a large group.Thus, high self-esteem was found to have one benefit, and itwas a benefit that fits well with theoretical assumptions. Takinginitiative and criticizing a group would seemingly require havingsome confidence in one’s own views, and sure enough, peoplewith high self-esteem were rated by their peers as morewilling and more likely to do those things. It is in fact surprisingthat the effect was so small.Group performance was also investigated in a longitudinalstudy by Paulhus (1998). He had 4-person and 5-person groupsmeet once per week for 7 weeks to discuss readings for theirpsychology class. Group members rated each other after thefirst and last weeks. Self-enhancement, defined as rating oneselfmore favorably than group members did, was positivelycorrelated with narcissism at the end of the study (whereas ifanything, narcissists came across as fairly modest at the initialassessment). In contrast, self-esteem showed a weak, but positivecorrelation with self-enhancement at both times. The mostrelevant outcome variable, however, was the rating of each individual’scontribution to group performance. People with highself-esteem were rated more positively than people with lowself-esteem at both times (r .25 and r .28, respectively),indicating good contribution to the group’s work. The heterogeneityof high self-esteem was evident, however, in the ratingsof narcissists. Initially, narcissists were rated quite positively ascontributing to the group, but by the end of the 7th week, thefavorable impression of narcissists had disappeared completely(r .01). In a second study, similar results were found, andin addition, narcissists went from generally positive peer ratingsinitially to being rated as cold, arrogant, inclined to brag,prone to overestimate self, and hostile by the end of the study.The second study also solicited ratings by friends prior tothe series of group meetings, which allowed Paulhus (1998) tocalculate a second, pretest, measure of self-enhancement, thatis, each person’s tendency to rate himself or herself more favorablythan friends did. The pattern of ratings for participantswith high scores on this pretest measure of self-enhancementwas quite similar to the pattern found for narcissists—peers’ratings of their contribution to the group were significantlypositive initially, but this positive impression was gone in thefinal ratings. The ratings at the end of the study depicted selfenhancersas confident, but cold, arrogant, prone to brag, defensive,and hostile.Leadership is another important aspect of group behavior,and most experts would say that the causes and determinants ofgood leadership form a set of questions that psychology has notyet managed to answer to everyone’s satisfaction. Intuitionsuggests that leadership ought to correlate with self-esteem, althoughthe causal arrow could point in either direction. Highself-esteem might give a person the confidence and initiative totake charge of a group, make decisions under pressure or uncertainty,and expect that others will do his or her bidding. Butit is also plausible that becoming a leader engenders a sense ofpersonal superiority and brings about a rise in self-esteem. Certainly,the careers of many exceptional leaders in history suggestthat they must have had some high degree of selfconfidencein order to make it to the top and that, conversely,VOL. 4, NO. 1, MAY 2003 19

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemover time their positions of power encouraged them to becomenarcissistic and sometimes megalomaniacal (Baumeister, 1989).An impressive study of leadership among military cadetswas recently published by Chemers, Watson, and May (2000).The cadets were followed over time, including during a specialsummer program devoted to leadership training. The measuresincluded ratings by peers and by military science professors, aswell as objective performance on two leadership tasks and twononleadership tasks (rifle marksmanship and land navigation).Simple correlational analyses showed some significant effectsof self-esteem. It predicted peer ratings, professors’ ratings,and performance on the second (though not the first) leadershipexercise, as well as performance on the two control nonleadershiptasks.Self-esteem, optimism, and leadership efficacy (feelings ofbeing a good leader) proved to be intercorrelated, however, sothe authors conducted more rigorous statistical analyses thatdisentangled the effects of these variables. In these analyses,global self-esteem did not have a significant correlation withleadership ratings obtained either from the professors or fromthe cadet peers. Self-esteem also failed to yield a significantcontribution to performance on the leadership exercises (as didoptimism), but leadership efficacy succeeded. The fact that theeffects of self-esteem disappeared in these more sophisticatedanalyses indicates that self-esteem overlapped with other factors(e.g., leadership efficacy) that may have taken precedencein predicting the outcome measures assessed. Chemers et al.found a unique association between self-esteem and performanceonly on the land navigation exercise. High self-esteemmight help people avoid getting lost (or, conversely, findingone’s way home might raise self-esteem).Group behavior also involves intergroup relations. In general,people favor their own groups in thought and deed. Theyprefer to assign greater rewards to members of their own groupthan to out-groups, they rate their own group more favorablythan out-groups, and they tend to direct prejudice and discriminationtoward out-groups. Early theorizing suggested that suchin-group favoritism is stronger among people with low selfesteemthan among those with high self-esteem (e.g., Ehrlich,1973). It seemed plausible that people low in self-esteem aremost inclined to boost themselves by means of disparaging anddiscriminating against members of other groups. Indeed, theopportunity to discriminate against a member of a negativelystereotyped group can bolster self-esteem (Fein & Spencer, 1997).Crocker et al. (1987) pointed out, however, that one cannotinfer the presence of prejudice when people with low selfesteemgive low ratings to out-groups. People with low selfesteemseem generally negative about many events and people,including themselves. A more appropriate measure of bias isthe difference between how one rates oneself or one’s own groupversus how one rates the out-group. By that measure, Crockeret al. found that prejudice was stronger among people withhigh rather than low self-esteem (see also Crocker & Luhtanen,1990).In a meta-analysis, Aberson et al. (2000) found that peoplehigh in self-esteem rate groups to which they belong more favorablythan they rate groups to which they do not belong. Peoplelow in self-esteem show no such difference. Recent workon social categorization suggests a simple process can accountfor this effect. When people predict the attributes of their ingroups,they rely heavily on their own self-concepts, but whenthey predict the attributes of out-groups, they do not engage insuch self-anchoring (or projection) (Cadinu & Rothbart, 1996;Clement & Krueger, 2002). Because positive attributes outnumbernegative ones in most people’s self-concepts, projectionsto in-groups yield positive outcomes. By not being subjectto projection, descriptions of out-groups remain more neutral.The process of differential projection also explains why ingroupbias typically reflects in-group favoritism rather thanout-group derogation (Brewer & Kramer, 1985).ConclusionPopular wisdom in the age of self-esteem holds that lovingoneself is a prerequisite for loving others (e.g., Crooks & Baur,1999). Efforts to boost self-esteem in schools, homes, and elsewherewould be well justified if they resulted in significant improvementsin how people got along with one another.The evidence suggests that the superior social skills and interpersonalsuccesses of people with high self-esteem existmainly in their own minds. People with high self-esteem claimto be more popular and socially skilled than others, but objectivemeasures generally fail to confirm this and in some casespoint in the opposite direction. When rated by peers, teachers,or laboratory interaction partners, people with high self-esteemare not liked any better than people with low self-esteem, andfollowing a blow to their pride they may become antagonisticand elicit negative reactions. People who have elevated or inflatedviews of themselves tend to alienate others.The sociometer model proposed by Leary and his colleagues(Leary et al., 1995) suggests that the level of self-esteem shouldbe an outcome of interpersonal successes and failures. This viewraises a very plausible alternative explanation for any findingsthat might link high self-esteem to social success. But as alreadynoted, there are relatively few such findings. Bishop andInderbitzen (1995) found that children without any friendstended to have low self-esteem, and it seems quite plausiblethat this lack of friends was a cause rather than a result of lowself-esteem.The results of studies on group behavior do not provide aringing endorsement of the importance of self-esteem. Highself-esteem does not appear to have much to contribute to leadership,especially after controlling for other variables. Highself-esteem appears to foster a small but significant tendencytoward greater voice behavior in work groups. This tendency tospeak up may also be responsible for people with high selfesteemreceiving somewhat higher ratings for contributing tothe group, compared with people with low self-esteem. People20 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.with high self-esteem also generally show greater bias in favorof their own groups, which may be beneficial for the in-groupbut may also produce greater discrimination and conflict betweengroups.Self-esteem does appear to have a variety of interpersonalconsequences. People with low self-esteem differ from thosewith high self-esteem in how they conduct their interpersonalrelationships and how they choose and pursue social goals.Still, these differences do not appear to translate into substantiallydifferent relationship outcomes.The most promising results involve social initiative. Thetendency to initiate interpersonal contacts and relationships (asrated by roommates) was the only interpersonal skill that differentiatedparticipants with high and low-esteem in the studyby Buhrmester et al. (1988). In groups, people with high selfesteemare more willing than those with low self-esteem tospeak up to criticize the group or propose directions for action.Likewise, people high in self-esteem take more initiative in extricatingthemselves from unhappy relationships (Rusbult et al.,1987). At present the evidence is sparse, but both theoreticalconsiderations and the few available findings suggest that selfesteemmay have value in promoting initiative.Self-esteem per se may not be the best variable for exploringinterpersonal processes. Narcissists (many of whom havehigh self-esteem) exhibit a variety of interpersonally problematicand alienating behaviors. Some depressed people (a groupthat includes some people with low self-esteem) constantlyseek interpersonal reassurance and alienate others for differentreasons.AGGRESSION, VIOLENCE, DELINQUENCY, ANDANTISOCIAL BEHAVIORFor decades, psychologists believed that low self-esteemwas an important cause of aggression. This view appears tohave emerged from clinical impressions rather than any singletheoretical formulation or line of empirical evidence. It waschallenged by Baumeister et al. (1996), who reviewed assortedbodies of findings and concluded that perpetrators of aggressiongenerally hold favorable and perhaps even inflated viewsof themselves. They proposed that aggression results fromwhat they called “threatened egotism,” namely, a positive viewof self that is questioned or attacked by other people. Twopoints need to be noted. First, that review did not contain laboratoryexperiments linking measured self-esteem to measuredaggression, and indeed many of the findings about violentgroups (such as incarcerated murderers) relied on interviewsand impressions rather than standard measures of self-esteem.Second, Baumeister et al. firmly embraced the notion of theheterogeneity of high self-esteem. They noted that many peoplehigh in self-esteem are not aggressive, even though othersare highly aggressive. If that view is correct, then carefullycontrolled and systematic studies attempting to measure bothself-esteem and aggression might generally find null results,insofar as the aggressive and nonaggressive people in the highself-esteem category would cancel each other out.AggressionAmong children, bullying is an important form of aggression.The traditional view that bullies have low self-esteem hasrecently come under attack. Olweus (1990, 1994) was one ofthe first to dispute the notion that bullies suffer from insecuritiesand self-doubts that underlie a tough exterior. Although Olweusdid not measure self-esteem directly, he did show thatbullies reported less anxiety and were more sure of themselvesthan other children. Analyzing self-reports of bullying behaviorfrom more than a thousand Australian schoolchildren, Slee andRigby (1993) found no correlation between bullying and selfesteem.Clearly, though, self-reports of bullying are suspect,insofar as people who score high on self-esteem generallyclaim to perform a broad range of socially desirable behaviorsmore often than other people do (and indeed they claimed toperform more prosocial behavior in this sample, too).Reliable information about bullying requires confirmationfrom external sources, preferably peers. Such data were obtainedby Salmivalli, Kaukiainen, Kaistaniemi, and Lagerspetz(1999), who had a sample of more than 300 Finnish adolescents(ages 14 and 15) fill out trait measures, including measuresof self-esteem and defensive egotism, and rate all theirclassmates on a variety of behavioral and trait measures. In additionto calculating correlations between self-esteem ratingsand other measures, Salmivalli et al. used the measures of selfratedand peer-rated self-esteem to categorize the participantsaccording to whether they were best characterized as havingdefensive high self-esteem, genuine high self-esteem, low selfesteem,a tendency to belittle themselves, or humble pride. Byusing these categories, the authors were able to allow for thepossible heterogeneity of high self-esteem.The correlation between each person’s self-esteem and theclassmates’ ratings of that person as prone to bully people wasnot significant. However, people categorized as having defensivehigh self-esteem (characterized by very high scores on defensiveegotism along with above-average scores on self-ratedand peer-rated self-esteem) were significantly more likely thanothers to be named by peers as bullies. Salmivalli et al. (1999)noted that this finding confirmed both the heterogeneity of highself-esteem and the threatened-egotism hypothesis put forwardby Baumeister et al. (1996). Of further interest was that the categoryof defensive high self-esteem covered not only the bulliesthemselves but their cronies, too. People in this categorywere rated by peers as prone to assist bullies in their victimizationof others, such as by joining in once a bully had begun attackingsomeone, or by catching and holding the victim so thebully could continue tormenting him or her. People with defensivehigh self-esteem were also rated as more likely than othersto reinforce bullying, such as by coming over to watch the bullying,shouting out encouragement to the bully, laughing at theVOL. 4, NO. 1, MAY 2003 21

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemspectacle, and making approving comments either to the bullyor to other bystanders.The heterogeneity of high self-esteem was confirmed bySalmivalli et al. (1999) in another way. Adolescents with genuine(as opposed to defensive) high self-esteem were rated bytheir peers as significantly more likely than others to defend thevictims of bullying. Upon reflection, this finding is not surprising,because coming to the aid of a victim requires some degreeof confidence. One might combine all these findings to suggestthat high self-esteem appears to support initiative, whether forgood or ill: People with high (defensive) self-esteem weremore likely than others to engage in bullying, and people withhigh (so-called genuine) self-esteem were more likely to fightagainst bullying. Meanwhile, people with low self-esteem weremore likely than others to be the victims of bullies.The work by Salmivalli et al. (1999) is the most rigorousand methodologically sophisticated study of bullying that wehave found, because it incorporated objective (peer-rated) measuresof bullying behavior and multiple measures of selfesteemthat allowed for heterogeneity in this construct. The maindrawback of this study is that it was cross-sectional and correlational,so that causality cannot be determined from it. Doesbullying other people cause (defensive) self-esteem to rise, ordoes defensive high self-esteem lead to bullying? The link betweenlow self-esteem and victimization can also quite plausiblybe construed in either causal direction: Bullies may choosevictims on the basis of perceived lack of self-esteem, or beingvictimized may be detrimental to self-esteem.Probably the best evidence about the causal links betweenaggression and self-esteem would come from programs specificallydesigned to alter self-esteem. A pair of studies of this typewas recently reported by Murphy, Stosny, and Morrel (2001),who tracked domestic offenders through therapy programs thatincluded interventions aimed at boosting self-esteem. Their researchhas several drawbacks that must be kept in mind. Onedrawback is that the sample consisted of men who were referredby courts to therapy on the basis of spouse abuse. It is generallybelieved that men in therapy for domestic violence differ fromother domestic offenders, not least because they know well thatthe situation demands that they express remorse and criticizethemselves. Another drawback of these studies is that the therapycould hardly focus on self-esteem alone, but also includedinterventions directly aimed at reducing violent behavior, andso one must be cautious in attributing any reduction in violentbehavior directly to increases in self-esteem.Nonetheless, the two studies are noteworthy because of theirfocus on real-world adult violence, and because in both cases,therapy led to significant increases in self-esteem. These increaseswere accompanied by significant self-reported reductionsin spouse abuse. Both before and after treatment, selfesteemcorrelated negatively with aggression, such that higherself-esteem was associated with less self-reported aggression,although the correlations reached significance in only one ofthe two studies.Self-reported violence is a problematic measure, however.Murphy et al. (2001) addressed this problem by obtaining reportsfrom the spouses and partners of the domestic offendersduring the 6 months after the therapy. These data alone meetthe standard of objectively confirming behavioral outcomes,and they provide no confirmation at all of the benefits of boostingself-esteem. There was no significant correlation betweenmen’s self-esteem or change in self-esteem and their partnerreportedabusive behavior, and if anything the trend was in thewrong direction: Higher post-therapy self-esteem and greaterincreases in self-esteem from the beginning to the end of therapywere both correlated with more rather than less subsequentspouse abuse (rs .08, n.s.).Another fairly rigorous approach to examining the impact ofself-esteem on aggression is to conduct laboratory experiments.Assuming causes precede effects, self-esteem measuredat the start of a laboratory session could not be the result of aggressivebehavior that is observed later in the session, althoughin principle a chronic tendency toward aggression might be theunderlying cause of both the aggression during the experimentand the self-esteem level. Baumeister et al. (1996) were unableto find any reports of laboratory experiments linking self-esteemto aggression. Given the wide popularity of both self-esteemand aggression as research topics, it did not seem plausible thatthe possible link between them had never been tested. Thispointed tentatively toward the conclusion that studies had beenconducted but remained unpublished because of null results.A pair of laboratory experiments by Bushman and Baumeister(1998) did provide a direct test of the role of self-esteem inaggressive behavior. The experiments investigated aggressionin direct retaliation for an insult, displaced aggression (i.e., participantswere insulted by one person and then had a chance toaggress against someone else), and unprovoked aggression(i.e., participants were praised rather than insulted and then hada chance to aggress against the praiser). The two studies useddifferent measures of self-esteem. Aggression was measuredby the loudness and duration of blasts of aversive, stressfulnoise participants delivered to an opponent when the opponentlost a trial in a competitive game.There were no significant correlations between levels ofself-esteem and aggression in any of the three situations studied.This confirmed the suspicion that the absence of previouslypublished findings probably indicated that studies hadbeen done but obtained null results.However, Bushman and Baumeister (1998) did find that narcissismwas a significant predictor of aggression. Specifically,narcissists were more aggressive than other individuals whenprovoked by an insult, but did not show elevated rates of displacedor unprovoked aggression. More recent studies by Stuckeand Sporer (2002) demonstrated high aggression among narcissistswho had been criticized and further suggested that thecombination of an unclear or confused self-concept with narcissismis especially likely to lead to aggressive retaliation. Theimpact of narcissism in those studies provided further support22 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.for the heterogeneity of high self-esteem, insofar as narcissistsmay be considered a category of high self-esteem people—people who not only think well of themselves, but also viewthemselves as being superior to others.DelinquencyJuvenile delinquency is of both theoretical and practical interest.It combines outright aggression, sometimes extending toassault and murder, with antisocial behavior such as stealingand vandalism. The assumption that low self-esteem causes aggressionhas traditionally offered one explanation for whyyoung people turn to delinquency, namely, that they lack selfesteem.However, evidence for the role of self-esteem in delinquencyis quite contradictory. In this section, we summarizesome of the more important findings.Neumark-Sztainer, Story, French, and Resnick (1997) obtainedself-reports of delinquent behavior by having adolescentsreport how frequently during the past 12 months they haddamaged or destroyed property, struck or beaten another person,or stolen from a store. Self-esteem showed a significantnegative correlation with self-reported delinquent behavioramong both boys (r .22) and girls (r .26), but in a regressionanalysis that controlled for family cohesion, school involvement,prior victimization, risk-taking disposition, andemotional well-being, self-esteem failed to emerge as a significantindependent predictor, despite the exceptionally largesample (N 123,132). The lack of independent predictiondoes not necessarily undermine the value of self-esteem: Selfesteemmay have important effects on delinquency that happento be mediated by other variables (e.g., perhaps low self-esteempeople are more likely to engage in delinquency because theyare more likely to disengage from school and take risks).A more recent study by Trzesniewski, Donnellan, Robins,Moffitt, and Caspi (2002, Study 1) with 292 fifth- and eighthgradestudents yielded significant correlations (rs ranged from.26 to .35) between self-reported delinquency and threedifferent measures of self-esteem (two self-report measuresand a teacher report). In this case, the relation remained significant(standardized coefficient .28) in a complex statisticalanalysis (latent structural equation model) that controlled forboth supportive parenting and academic achievement.One of the most ambitious and intriguing studies of selfesteemand delinquency was published by Rosenberg et al.(1989). Their nationwide sample included nearly 2,000 adolescentmales who were measured twice. The study’s longitudinaldesign permits some causal inference, although its exclusivereliance on self-reports of delinquent behavior means that theresults may be compromised by self-report biases. The simplecorrelations between self-esteem and self-reported delinquencywere quite weak in this study, rs .09 and .07 at the firstand second assessments, respectively. Rosenberg et al. proposed,however, that this seemingly null result might actuallyconceal contrary trends. By examining changes across time,they concluded that there were actually two significant linksbetween self-esteem and delinquency in the data. First, lowself-esteem was associated with an increase in delinquent behavior(r .19). Second, delinquent behavior was associatedwith an increase in self-esteem (r .08, marginally significant).They also found that these effects interacted with socialclass. Low self-esteem was most prone to increase delinquencyin the more affluent group, possibly because delinquent behaviorwas more normative among the lower classes, so the numberswere already so high that there was not much room forincrease. Meanwhile, the boost in self-esteem deriving fromdelinquent participation was greater among the young men oflow rather than high (or medium) socioeconomic status.Another recent longitudinal study, however, failed to confirmthe conclusions of Rosenberg et al. (1989). In this study, Jangand Thornberry (1998) followed nearly a thousand students(starting in seventh and eighth grade) for 4 1/2 years. They distinguishedbetween delinquent activities and associating with delinquentfriends, although self-reports were the main source ofinformation about both. Jang and Thornberry found no tendencyfor low self-esteem to predict subsequent delinquency, whethermeasured by criminal and antisocial behavior or having delinquentfriends. They also failed to find that delinquent activityboosted self-esteem, and, if anything, self-reported delinquentactivity led to a decrease in self-esteem. Forming friendshipswith delinquent peers did, however, boost self-esteem. Thus, anyapparent benefits from delinquency appeared to be due to formingfriendships and engaging in social comparison with delinquentpeers, as opposed to committing violent or antisocial acts.Most recently, an ambitious longitudinal study by Trzesniewskiet al. (2002, Study 2) yielded fairly strong evidencethat self-esteem may indeed play a causal role in externalizingbehavior, a constellation of behaviors that includes delinquencyplus additional antisocial misbehavior not generally considereddelinquent. (The term externalizing behavior reflects the assumptionthat the person is dumping his or her problem ontoothers.) This study followed students from age 11 (n 809;48% female) to age 13 (n 726; 48% female) and assessed externalizingbehavior with three relatively objective measures:teacher report, parent report, and an interview report (at age 13only) based on the American Psychiatric Association’s symptomcriteria for Conduct Disorder. Self-esteem was significantlycorrelated with externalizing behavior at both time points (rsranged from .16 to .25), and self-esteem at age 11 predictedexternalizing behavior at age 13 (rs ranged from .19 to.21). Moreover, the contribution of self-esteem was found tobe independent of relationships to parents, relationships withpeers, IQ, and socioeconomic status. Given the large sample,longitudinal design, and multimethod measurement of externalizingbehavior, this study provides the best available evidencefor a positive link between low self-esteem and subsequent delinquentbehavior.Overall, there is some support for the traditional view thatlow self-esteem may predispose a person to participate in anti-VOL. 4, NO. 1, MAY 2003 23

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemsocial behavior. The correlation between self-esteem and delinquencyvaries among studies (from nearly zero to around.30), but it is almost always negative, suggesting that the effectis present, although probably quite weak. There is alsosome suggestion that the effect of self-esteem on delinquencymay be mediated by variables such as school involvement andchoosing to associate with delinquent friends.Antisocial BehaviorAntisocial behavior is a companion to delinquency. By antisocialbehavior, we mean behavior that is socially undesirablebut not necessarily violent. An experiment by Lobel and Levanon(1988) examined cheating in relation to individual differencesin self-esteem and need for approval. The participantswere 12-year-old boys and girls in school in Israel. Their taskwas to trace line drawings without lifting the pencil or retracinglines. After two easy problems, three unsolvable problemswere presented, and the children could claim to have solvedthem only if they cheated, which thereby allowed the researchersto obtain an objective measure of cheating. Self-esteem didnot have a direct effect on cheating, but interacted with needfor approval to determine the amount of cheating. The resultswere consistent with the view that self-esteem is a heterogeneousconstruct, in that the children who cheated the most andwho cheated the least both had high self-esteem. More precisely,those who were high in both self-esteem and need forapproval cheated the most, whereas those who were high inself-esteem but had a low need for approval cheated least. Theauthors proposed that a distinction between true and defensivehigh self-esteem could account for their results.School disciplinary problems and related antisocial behaviorswere the focus of a large survey by DeWit et al. (2000). Allstudents in four secondary schools in Ontario, Canada, were includedin the study, for a total of a thousand participants. Althoughthe authors clearly favored the self-esteem hypothesis,self-esteem was not a significant predictor for three of the fiveoutcome variables. Most notably, self-esteem did not contributeto the prediction of disciplinary problems in school, substanceabuse, or conduct problems.DeWit et al. (2000) relied exclusively on self-reports of antisocialbehavior. As we noted earlier, all self-reports are notequal, with reports of concrete and verifiable misbehavior presumablybeing less subject to distortion and bias than global estimates.The reports of school disciplinary problems seem themost concrete and verifiable measure DeWit et al. used, insofaras students were asked how frequently they had been suspended,served detentions, been sent to the principal, or beentruant in the past 6 weeks. The measures of substance abuselikewise seem reasonably concrete, although less readily verifiable,and on these measures self-esteem yielded no significanteffects. In contrast, oppositional-defiant disorder (a recurringpattern of negative, hostile, disobedient, and defiant behavior)was related to low self-esteem. Indications of this disorder weremeasured largely by an attitude survey consisting of questionssuch as, “How often in general do you get annoyed by others?”The fact that self-esteem had its effects on the less concretemeasures suggests some potential role for self-report bias.Last, a set of interview and laboratory studies by Colvin,Block, and Funder (1995) looked specifically at people with inflatedself-esteem. This was done by comparing how favorablypeople rated themselves with how favorably they were rated byothers, and the people whose self-ratings were higher than thepeer ratings were classified as having inflated favorable viewsof self. Participants were then invited into the laboratory, andtheir interactions with randomly assigned peers were observed.As compared with controls, people who had inflated self-esteemexhibited a high rate of antisocial and socially undesirablebehaviors. They were more prone to interrupt othersduring the conversation. They expressed more hostility andwere found to annoy and irritate their partners. Their conversationalstyle was seen as talking at rather than talking withsomeone. Despite their favorable self-ratings, they came acrossas socially awkward. In a second study, people of this sort werefound to be prone to psychological maladjustment and to havepoor social skills.ConclusionMany researchers have sought to link self-esteem to violence,aggression, and antisocial tendencies. The results aremixed at best. Psychologists who wish to study or reduce aggressionmight be well advised to focus on factors other thanself-esteem or, at least, to respect the heterogeneity of highself-esteem and therefore consider additional variables. In thecase of aggression, simple measures of self-esteem have generallyfailed to predict objective behavior, whereas high narcissism(presumably one subcategory of high self-esteem) doeslead to aggressive retaliation, a finding consistent with thethreatened-egotism model. Inflated self-esteem, in the sense ofregarding oneself more favorably than one’s peers do, is alsoassociated with socially undesirable behaviors. Although mostlarge studies have found self-esteem to be weak at predictingdelinquency, school disciplinary problems, and related antisocialactivities, and in fact some findings suggest that peoplewith some varieties of high self-esteem are most prone to actantisocially, one methodologically strong recent study has suggestedthat low self-esteem may predispose individuals towardexternalizing behavior.Although further research is needed, one impression thatemerges from these data is that self-esteem simply intensifiesboth prosocial and antisocial tendencies. As the study by Salmivalliet al. (1999) showed, people who think well of themselvesare overrepresented among both the perpetrators ofbullying and the people who stand up to bullies and defend victims.By the same token, Lobel and Levanon (1988) found thatpeople high in self-esteem made up both the highest-cheatingand the lowest-cheating groups. Quite possibly, the actual ef-24 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.fect of high self-esteem per se is to support initiative and confidentaction, for good or ill.HAPPINESS, COPING, DEPRESSIONWe turn now to subjective outcomes, such as happiness. Thepursuit of happiness is one of the overarching goals in life formost people, and feeling happy is one form of positive outcomewhose desirability is difficult to dispute. Along with happiness,we consider two other related phenomena. One of theseis depression, which is often defined (at least in part) by pervasiveunhappiness. The other is coping with stress and trauma.Even if self-esteem were to fail to contribute to a better, happierlife overall, it might accomplish a great deal by helpingpeople to recover from aversive events. Misfortunes, problems,and setbacks can be experienced as blows to one’s favorableview of self (e.g., Janoff-Bulman, 1992; Taylor, 1983), and it isplausible that high self-esteem enables people to recover fasteror more thoroughly from such events.Before we present findings, we must acknowledge that studiesof happiness and related variables seem almost inevitably to relyon self-reports. For other outcomes, we were routinely suspiciousand critical of self-report data, and whenever possible we soughtobjective measures. Yet it is not clear what sort of objective measurecould replace self-reports of happiness. It is possible to provethat people are less (or more) intelligent, attractive, or likable thanthey think they are, but in what sense could researchers concludethat people are less (or more) happy than they think they are? Ifsomeone estimates his or her happiness to be at 60 on a scale runningfrom 1 (very low happiness) to 100 (very high happiness),there is no objective criterion, or “happiness meter,” that couldshow that the person’s true happiness is higher or lower. There isnot even any meaningful or realistic way to assert that someonecan be mistaken about his or her happiness. As a result, studies investigatingany possible link between self-esteem and happinesswill necessarily have to rely on self-reports for both measures, despitethe pitfalls and drawbacks of self-reports in connection withestablishing the effects of self-esteem.Reliance on self-report carries the danger of distorting therole of self-esteem. As we reported in preceding sections, peoplewith high self-esteem repeatedly emerge as better than othersin self-report data, whereas objective data often depict suchpeople to be no better than others. A response bias toward presentingthe self in favorable, socially desirable ways appears tocharacterize people high in self-esteem. Hence, one might wellexpect that people with high self-esteem claim to be happierthan other people. However, there does not seem to be any wayaround this problem, given the lack of objective measures (indeedthe ineluctable subjectivity) of happiness.HappinessA major international study of self-esteem and happinesswas reported by Diener and Diener (1995). The data came frommore than 13,000 college students from 49 different universities,31 countries, and five continents. High self-esteem emergedas the strongest of several predictors of life satisfaction overall.The simple correlation between self-esteem and happiness wasquite significant at .47. In short, self-esteem and happiness aresubstantially interrelated.The correlation between self-esteem and happiness (life satisfaction)varied somewhat across countries. In particular, itwas stronger in individualistic countries than in collectivisticones. Diener and Diener (1995) proposed that in individualisticcountries, people are socialized to attend to their own internalattributes, and so these become important in determining overallhappiness, whereas collectivistic cultures encourage peopleto focus on groups and relationships.In a sample of 406 young people (aged 14–28), Furnhamand Cheng (2000) measured a number of potential correlates ofhappiness, including recalled parental rearing styles, scores onthe Eysenck Personality Questionnaire, and self-esteem. Theyreported that self-esteem was the most dominant and powerfulpredictor of happiness. Shackelford (2001) examined the happinessof young to middle-aged couples (ages ranged from 17to 41) who had been married within the past year and foundthat for both husbands and wives, self-esteem was significantlycorrelated with happiness in the form of global, sexual, andemotional satisfaction.Data on older adults were provided recently by Lyubomirskyand Lepper (2002), and these nicely complement the emphasison college students in Diener and Diener’s (1995) study.Lyubomirsky and Lepper obtained data from more than 600adults ranging in age from 51 to 95. Once again, happiness andself-esteem were highly correlated (r .58). It is important tonote that in both this study and Diener and Diener’s work, selfesteemand happiness had very different patterns of correlationswith other predictor variables, which supports the conclusionthat happiness and self-esteem are in fact distinctconstructs that can be measured separately despite their relativelyhigh intercorrelation. In Lyubomirsky and Lepper’sstudy, self-esteem was more strongly correlated than happinesswith hopelessness, optimism, and sense of mastery, whereashappiness was more strongly correlated than self-esteem withenergy level, overall health, loneliness, mood and emotion, andpurpose in life.A meta-analysis of the relation between 137 personalitytraits and happiness (subjective well-being) was published byDeNeve and Cooper (1998). They found that “private collectiveself-esteem” (a term they did not explain) was one of thestrongest predictors of happiness (r .31). Other forms of collectiveself-esteem yielded weaker results in their analysis.They do not appear to have included the more typical measureof self-esteem, nor did they define their terms, but in any casetheir results provide further evidence that self-esteem (of somesort) is consistently correlated with happiness.Taken together, these findings uniformly indicate that selfesteemand happiness are strongly interrelated. They suggestVOL. 4, NO. 1, MAY 2003 25

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemthat high self-esteem may pay off handsomely for the individualin terms of subjective happiness. Our main note of cautionis that the primitive data resemble the simple and previouslypromising correlations between self-esteem and school achievement.Further research with longitudinal designs (or other meansof testing causal relationships), controls for third variables, and,if possible, alternatives to self-report measures would greatlystrengthen the case.Coping and DepressionStudies of coping look specifically at how people respond toharmful, threatening, traumatic, or otherwise unpleasant events.The hypothesis favored by most theorists is that high selfesteemoperates as a kind of resource that enables people tosuffer less or to recover more rapidly from such events (e.g.,Arndt & Goldenberg, 2002). According to this hypothesis, underlow stress, people would fare reasonably well regardlessof self-esteem, but in the wake of stress, people with high selfesteemwould fare better than those with low self-esteem.Sometimes this hypothesis is labeled the buffer hypothesis, becauseit asserts that high self-esteem operates as a bufferagainst stress, trauma, and misfortune.To test the buffer hypothesis, Murrell, Meeks, and Walker(1991) interviewed a sample of more than a thousand adults at6-month intervals for several years. Contrary to their predictions,self-esteem failed to operate as a buffer, in that the effectof self-esteem on depression was essentially unchanged whetherstressful events were included in the statistical analysis or not.Instead, high self-esteem predicted lower depression overall,regardless of stress. In other words, low self-esteem predisposedindividuals toward depression in both good times andbad, and actual circumstances made little difference.A similar finding was obtained by Robinson, Garber, andHilsman (1995). Their original sample was composed of nearly400 sixth-grade students, almost 300 of whom were interviewedagain in the fall of seventh grade (and more than 200were surveyed again in the spring of that year). This timing isof particular interest because many students move from elementaryto junior high schools between sixth and seventhgrade, and the change of school can be a potent source ofstress. Self-esteem had a direct effect on depression but not onexternalizing behaviors. There was no interaction between selfesteemand stress in predicting either dependent variable. Thus,as Murrell et al. (1991) found, low self-esteem predicted depressionregardless of life stress.It must be noted that Robinson et al. (1995) did find a threewayinteraction between stress, self-esteem, and attributionalstyle. That is, depression was increased by a combination ofexternal stresses and a tendency to blame oneself for failure,especially among people low in self-esteem. Put another way,low self-esteem alone did not create a vulnerability to becomingdepressed in response to stress, but low self-esteem combinedwith a pessimistic and self-blaming style of thinking did.Another pair of studies did find interactions between selfesteemand life stress predicted participants’ increases in depressivefeelings. Whisman and Kwon (1993) tested a sampleof 80 undergraduates on two occasions separated by 3 months.They found that increases in dysphoria (a term for depressionthat is sometimes preferred because it avoids the connotation ofclinical pathology) were predicted by self-esteem, by lifestress, and by the interaction between the two. As in the studiesalready discussed, low self-esteem led to greater depression.Unlike in those studies, high life stress also led to greater depression,and the interaction between stress and self-esteemwas also a significant predictor.However, the interaction Whisman and Kwon (1993) founddid not conform to the pattern of the buffer hypothesis. Thebuffer hypothesis holds that high self-esteem helps people copewith life stress, so differences between people with high andlow self-esteem should be found mainly under conditions ofhigh life stress. Instead, Whisman and Kwon found the opposite:The self-esteem differences were most pronounced underconditions of low life stress. Moreover, Whisman and Kwonreported that life stress appeared to have the biggest effect onpeople with high (rather than low) self-esteem. That is, peoplewith high self-esteem were happy in good times but unhappyduring stressful times, whereas the degree of life stress apparentlymade less difference to people low in self-esteem.A larger study along the same lines was conducted by Ralphand Mineka (1998), who surveyed a sample of 141 undergraduatesmore than a week before their midterm examination,again after the exam but before grades were issued, the day afterthe students received their grades, and 3 days later. Thisprocedure allowed the researchers to track reactions to a particular,well-defined stressful experience with a positive or a negativeoutcome. The distress of people with high self-esteemtended to decrease after they received their grades, regardlessof what the grades were. In contrast, the responses of low selfesteempeople depended on how their grades compared withwhat they had sought and expected. (The researchers had askedthe students to indicate the lowest acceptable grade, and the relationshipbetween that answer and the actual grade was usedto classify each student’s outcome as a personal success or failure.)Contrary to the buffer hypothesis, the greatest increases ingeneralized distress occurred among people who had low selfesteem(and also high pessimism) and received grades thatwere acceptable to them. Thus, it was the reaction of peoplewith low self-esteem to a relatively good outcome that wasmost problematic and revealing.The studies by Whisman and Kwon (1993) and Ralph andMineka (1998) produced findings that contradicted the bufferhypothesis: Differences between people with high versus lowself-esteem emerged under relatively positive, benign conditions,rather than under stressful conditions. Moreover, lowself-esteem seems to poison the good times. In contrast, thebuffer hypothesis did receive support from DeLongis, Folkman,and Lazarus (1988) in a study of 75 married couples as-26 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.sessed 20 times over 6 months. They measured daily hassles,physical health and symptoms, and mood. Perhaps surprisingly,self-esteem did not moderate the impact of stressful hassleson mood, but it did moderate the link between stressfulhassles and physical symptoms. Participants who were low inself-esteem showed a stronger link than others between theamount of stress they experienced on a particular day andwhether they had physical illness on the same day and on thefollowing day. Moreover, the buffering effect of self-esteem remainedsignificant even in analyses that controlled for the participants’emotional support and the size of their socialnetworks. Thus, people with low self-esteem were apparentlymore prone than others to get sick or suffer other physicalproblems in connection with stressful daily events.Metalsky, Joiner, Hardin, and Abramson (1993) also foundsupport for the buffer theory, although only under certain conditions.They assessed self-esteem and other factors in a sampleof more than a hundred students prior to midterm gradereports and then took several measures on the day midtermgrades were received and for 5 consecutive days thereafter.Subjective failure on the midterm was defined as getting agrade below the level the student had previously stated as acceptable.Increases in depression at first depended solely onsubjective failure, regardless of other factors. Over time, however,depressive reactions were found to be a product of athree-way interaction. Substantial increases in depression werefound only among people who fit three criteria: subjective failure,negative (self-blaming) attributional style, and low selfesteem.(These results are quite consistent with the pattern offindings reported by Robinson et al., 1995, described earlier inthis section.) Thus, low self-esteem, especially in combinationwith a self-blaming attributional style, made people more vulnerableto feeling depressed in response to a real-life stressorover a short period of time.A similar investigation by Roberts and Monroe (1992)reached a somewhat different conclusion. They, too, tracked asample (N 216) of college students, and measured self-esteembefore midterm exams and depression afterward. Academicstress was assessed by calculating the difference between a student’shoped-for grade and actual grade, and then multiplyingthis difference by the subjective importance (to the student) ofacademic achievement. There was a small independent effectof self-esteem on depression, such that students with lowerself-esteem just before the examination reported more depressionafter getting their grades, regardless of what their gradeswere. This effect is interesting, despite its small size, because itshows low self-esteem leading to increased depression over ashort period of time. Unfortunately, the effect was not maintainedin the later days of the study, which is puzzling and castsdoubt on the impact of self-esteem on depression.Level of self-esteem did not interact with academic stress topredict depression in Roberts and Monroe’s (1992) study, contraryto the buffer theory. However, unstable self-esteem (measuredin terms of within-person variance in self-esteem scoresduring the preexamination, baseline period) did. Depression increasedmost among students who were initially not depressed,had unstable self-esteem, and subjectively failed on the examination.It also increased among those who were initially depressed,had unstable self-esteem, and did well on the exam.Unstable self-esteem also independently predicted an increasein depression regardless of other factors, although when statisticalcontrols were entered into a more complex analysis, thiseffect vanished.Building on this work and other findings, Roberts and Monroe(1994) proposed a general theoretical account of the role ofself-esteem in depression. They acknowledged that low selfesteemhas often been proposed as a risk factor that creates avulnerability to depression, but concluded that in research,level of self-esteem has failed to emerge as a robust predictorof the onset of depression. They proposed that vulnerability todepression accompanies unstable self-esteem (i.e., self-esteemthat is prone to fluctuate across time), as well as self-esteembased on relatively few and unreliable sources.The buffer theory was supported in Corning’s (2002) researchon reactions to perceived discrimination among women.As in most other studies, there was a general tendency for peoplewith high self-esteem to feel better than those with low selfesteemregardless of events. However, self-esteem interacted withperceived discrimination. Women with low self-esteem reactedto being the target of discrimination by becoming distressed,whereas the effects of discrimination were much weaker onwomen with high self-esteem. Corning examined a variety ofbad outcomes, including anxiety, depression, and physical symptoms,rather than focusing exclusively on distress.Recent work by Roese and Pennington (2002) suggests apossible reason why people with high self-esteem apparentlycope better with discrimination than people with low selfesteemdo. People with high self-esteem tend to have a gap betweentheir perception of discrimination against their groupand their perception of discrimination against themselves. Thatis, they believe that their group is often the victim of discrimination,but that they themselves have not been victims.Laboratory experiments by Greenberg et al. (1992) alsosupported the buffering hypothesis, although with regard toanxiety rather than depression. They sought to boost self-esteemby first giving people a personality test and then providing bogusfeedback that conveyed a very positive evaluation. Afterward,the participants were exposed to one of several stressors,including having to watch upsetting images of gruesome deathand receiving bad feedback on an IQ test. The stressors tendedto produce increases in self-reported anxiety and signs of physicaldistress, but these effects were attenuated among peoplewhose self-esteem had been boosted.The buffer hypothesis also received support from a recentpair of studies by Bonanno et al. (2002), although they used ameasure of self-enhancement (i.e., whether participants ratedthemselves as better than average and more highly than peersrated them on various dimensions) rather than a standard self-VOL. 4, NO. 1, MAY 2003 27

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemesteem measure. Bonanno et al. examined reactions to stress intwo very different populations, residents of Bosnia during therecent civil war and young to middle-aged residents of the SanFrancisco Bay area whose spouse had died within 3 years ofmarriage. The full buffer hypothesis could not be tested becausethese studies did not include low-stress control groups,but in both studies, self-enhancers reported better coping thannon-self-enhancers. Moreover, the findings for the self-reportdata were supported by more objective data: The good adjustmentof the self-enhancers was confirmed in mental health experts’ratings based on clinical interviews.How can all these findings on the relation between depressionand self-esteem be integrated? First, nearly every studydid find that people low in self-esteem felt more depressed orotherwise worse than people high in self-esteem. Second, thebuffer hypothesis (namely, the hypothesis that low self-esteemcreates vulnerability to stress) received its best support in studiesthat used outcome measures other than depression, such asphysical illness or anxiety. Third, when low self-esteem did interactwith life stress, the findings were about evenly split as towhether they supported the buffer hypothesis or the oppositepattern—that is, half the studies confirmed that low self-esteemincreased vulnerability to stress but was irrelevant in goodtimes, whereas the rest found that low self-esteem poisoned thegood times rather than exacerbating vulnerability to the badtimes. Last, some findings favor a more complicated version ofthe buffer hypothesis that incorporates a third variable, namely,self-blame. In other words, the combination of low self-esteem,self-blame, and stressful or aversive events causes depressionto increase.ConclusionThe link between self-esteem and happiness is strong. Peoplewith high self-esteem are significantly, substantially happierthan other people. They are also less likely to be depressed,either in general or specifically in response to stressful, traumaticevents. Many studies have confirmed this link.Before it is safe to conclude that self-esteem causes happiness,however, further research must address the methodologicalshortcomings of the work that has been done so far. First,causation needs to be established, and potential third-variablecauses need to be ruled out. It seems plausible that self-esteemcauses happiness, but no research has shown this. The strongcorrelation between self-esteem and happiness is just that—acorrelation. It is possible that occupational, academic, or interpersonalsuccesses cause both happiness and high self-esteem,whereas corresponding failures cause both unhappiness andlow self-esteem. It is even possible that happiness, in the senseof a temperament or disposition to feel good, causes high selfesteem.In this regard, the research on academic performance isinformative: The correlations between self-esteem and goodgrades do not seem to derive from any causal impact of selfesteemon schoolwork, but rather other background causes leadto both, and to some extent good school performance leads tohigher self-esteem. Happiness seems to us to be more promisingthan academic performance as a potential consequence ofself-esteem, but the alternative causal pathways need to be consideredwith equal rigor and care.Second, happiness and depression have been studied mainlyby means of self-report. As we have seen, the tendency of somepeople toward general floccinaucinihilipilification may produceboth their low self-esteem and unfavorable reports aboutother aspects of life. Clearly, objective measures are muchmore difficult to obtain for happiness and depression than foracademic performance. But the lessons from academic performance,interpersonal skill, physical attractiveness, and the likeare sobering with regard to self-reports.Support for the buffer hypothesis is uneven, but there areseveral strong and supportive findings. The correlations betweenhigh self-esteem and happiness and between low self-esteemand depression may conceal some influence of circumstances.Yet further research is needed to untangle these processes.Some studies supported the buffer hypothesis: Low self-esteemproduced its detrimental effects mainly in bad times. Othersfound the opposite: Low self-esteem seemed to poison thegood times, whereas in bad times everyone suffered. Still othersfound circumstances to be irrelevant, and low self-esteemsimply predicted worse outcomes in both good times and bad.The final resolution of this theoretical controversy is not insight, and we recommend that researchers continue to conducttests and analyses to illuminate this question, all the while beingcognizant of conflicting findings. It seems fair to assumethat the ultimate answer will be more complex than the originalbuffer hypothesis. Nonetheless, the findings consistently suggestthat low self-esteem leads to poorer outcomes, includingdepression and possibly physical illness, under some circumstances.HEALTH, SMOKING, ALCOHOL AND OTHERDRUGS, SEX, EATINGThere are a number of reasons to believe that high levels ofself-esteem may have a beneficial impact on physical health.For example, as we noted earlier, high self-esteem individualsperceive that they have more social support, and perceptions ofsocial support have well-documented relations with numeroushealth outcomes, including length of life (see Ornish, 1998, fora summary).Another reason to surmise that high self-esteem might havebenefits for health lies in the vast literature demonstrating theimpact of stress (especially chronic stress) on health. Stress is,of course, a highly subjective response, and depends criticallyon both the individual’s interpretation of events and his or herbeliefs about coping resources that are available (Lazarus &Folkman, 1984). Because high self-esteem people interpret28 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.negative events more benignly, perceive that they have greatercontrol, and are generally more optimistic about their copingabilities, it seems reasonable that they would experience lessstress and therefore enjoy better levels of physical health thanpeople with low self-esteem.The adverse health consequences of stress are primarily mediatedby activation of stress hormones (cortisol). Thus, thebest indirect evidence for the supposition that high self-esteembenefits health comes from studies demonstrating that peoplewho differ in self-esteem show differential cortisol reactivity inresponse to stressors experienced in the laboratory. In general,these studies have shown that when the stressor involves actualfailure or the strong potential for failure (e.g., for elderly people,a challenging test of ability to drive a car), cortisol responsesare higher in participants with low self-esteem than inthose with high self-esteem (Kirschbaum et al., 1995; Pruessner,Hellhammer, & Kirschbaum, 1999; Seeman et al., 1995).Nonetheless, there are only a few studies that provide anydirect evidence linking self-esteem and objective health outcomes.In a study of 502 adults who were over 60 years of ageand had cardiovascular disease, high self-esteem predicted betterphysical and psychosocial functioning 12 months lateramong female, but not male, participants (Forthofer, Janz,Dodge, & Clark, 2001). And an earlier study of more than1,300 Helsinki policemen showed that among those whoseelectrocardiograms had signs of coronary heart disease, lowerself-esteem was related to a higher risk of dying from myocardialinfarction over the next 10 years (Nirkko, Lauroma, Siltanen,Tuominen, & Vanhala, 1982).Several studies have found that people high in self-esteemrate their overall health better than people with low self-esteem(e.g., Glendinning, 1998). James (1997) found that collectiveself-esteem (such as racial pride) predicted self-reported healthin a small sample of African American workers but not Whiteworkers. In a large sample of 1,700 Scottish youth, aged 14 to 16,Glendinning (1998) found that self-esteem was linked to selfratingsof health, but not to self-reports of health-relevant behaviors.Still, these latter findings were not verified by objectivemeasures, and could be distorted by the usual self-report bias.Although we regard it as premature to draw any conclusionsabout the impact of self-esteem on physical health, this doesseem an area that deserves further study (and any null findingsshould be published). In particular, further investigation throughlongitudinal designs that control for potential mediators ormoderators of the relation would be desirable. It is conceivablethat the benefits of high self-esteem, including feeling good,lead directly to better health. Then again, the studies we surveyedall found self-esteem to be correlated with better healthonly in some groups and not in others, and they differed as towhich groups benefited. It is also possible that underlying biologicalcauses, such as temperament or good physical condition,contribute to both self-esteem and health, in which case interventionsaimed at raising self-esteem would likely have no effecton health. In any case, more and better research is warranted.Unlike overall health, many health-relevant behaviors havebeen studied in relation to self-esteem. Foremost among theseare several sets of destructive or risky behaviors that youngpeople engage in. The public interest in self-esteem is motivatedin part by the hope that boosting self-esteem will helpyoung people abstain from abusing alcohol and other drugs,smoking cigarettes, succumbing to pathological eating patterns,and engaging in risky sexual practices. We summarizethe findings regarding these behaviors in this section.A major shortcoming of most of these studies is that theyrely primarily on self-reports. To be sure, it would be neitherethical nor practical to attempt to induce youngsters to take illegaldrugs or engage in risky sex during a laboratory experiment,and attempting to observe these covert activities in theworld outside the laboratory is also not very practical. Hence,some degree of reliance on self-report is difficult to avoid. Still,interpretation of the findings should be tempered with recognitionthat people high in self-esteem will tend to report more sociallydesirable patterns of behavior than people with low selfesteem.Also, self-reporting of specific behaviors is presumablymore reliable than vague, general self-ratings.SmokingSmoking is widely recognized as detrimental to health. Ifhigh self-esteem could help prevent smoking, even just amongyoung people, that would be a valuable contribution to individualwelfare and society at large. It is also possible that smokingaffects self-esteem rather than the reverse, because the selfesteemof smokers who feel stigmatized may suffer. Several largeinvestigations have explored possible links between self-esteemand smoking, although smoking has generally been measuredonly by self-report, and findings that people with high self-esteemsmoke less than others might be an artifact of self-report bias.Three studies have suggested that low self-esteem is linkedto smoking, but unfortunately none of them shows that lowself-esteem clearly leads to subsequent smoking. Pederson,Koval, McGrady, and Tyas (1998) surveyed more than 1,600Canadian eighth graders, distinguishing multiple categories ofsmokers. The current smokers had lower self-esteem than theadolescents who had never smoked. The self-esteem of exsmokerswas in between, and experimental smokers (i.e., thosewho tried out smoking occasionally without becoming regularsmokers) had nearly the same level of self-esteem as the eighthgraders who had never smoked.In a younger sample of nearly 1,500 third-grade and fifthgradestudents in central North Carolina, Jackson, Henriksen,Dickinson, and Levine (1997) found that children with lowerself-esteem were more likely to admit to having smoked on occasion.A longitudinal study by Andrews and Duncan (1997)tracked more than 400 adolescents, whose age at the start ofthe study was between 11 and 15, for 3 years. The researchersVOL. 4, NO. 1, MAY 2003 29

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemwere particularly interested in whether self-esteem and othervariables would mediate the relation between academic motivationand cigarette smoking (among other variables). Selfesteemdid not mediate this relation, but it was an independentpredictor of smoking. When self-esteem and smoking were measuredat the same time, the average correlation between themwas .18. Lagged correlations across time yielded no clearpattern as to whether self-esteem predicted subsequent smokingor smoking predicted subsequent self-esteem.Two additional studies have found that the link betweenself-esteem and smoking depends on gender. Lewis, Harrell,Bradley, and Deng (2001) surveyed 1,200 10- to 15-year-oldsliving in three tobacco-producing counties in North Carolina.Among girls, smokers had significantly lower self-esteem thannonsmokers. Among boys, the smoking and nonsmokinggroups did not differ in self-esteem. A sample of more than8,000 children (which shrank to 6,530 by the 4th year of thestudy) in Calgary, Alberta, Canada, was studied by Abernathy,Massad, and Romano-Dwyer (1995). These researchers foundthat for girls, low self-esteem in grade 6 predicted a substantiallygreater likelihood of smoking by grade 9. Indeed, girlswith low self-esteem were about 3 times more likely to try cigarettesthan other girls. Among boys, self-esteem had no relationshipto smoking. This study is notable both for its largesample and for its prospective, longitudinal design.In contrast, several large studies found no relationship whatsoever.The possible link between self-esteem and smokingwas the specific and primary focus of research by Glendinningand Inglis (1999). They reported findings from two Scottishsurveys (one national, one local), which included a total ofnearly 3,000 young people aged 13 and 14. Noting weaknessand inconsistency in past findings, they included several measuresof smoking, such as frequency of smoking and havinggiven up smoking versus never having smoked. They alsosorted their respondents into various social categories, such associally isolated, conventional, and peer oriented. Despitethese efforts, the data showed no significant relation betweenself-esteem and smoking, even when social categories weretaken into consideration.A longitudinal study of a thousand young people in NewZealand was reported by McGee and Williams (2000). The researchersassessed self-esteem at age 12 and then inquiredabout an assortment of health-compromising behaviors at age15. There was a slight trend toward greater smoking among individualslow in self-esteem, but it failed to reach significance,and given the large sample, a lack of significance must be takenas an indication that the relation was negligible.As part of a prospective, longitudinal study of health patternsin a cohort of a thousand young Scots, West and Sweeting(1997) reported on the participants when they were 15. Therewas a slight tendency for the young people with the highestself-esteem to report less smoking than others, but this failed toreach significance despite the large sample. In another largesample (1,700) of rural Scottish youth, ages 14 and 16,Glendinning (1998) found that self-esteem had no correlationwith self-reported smoking.A Canadian sample of more than 1,500 sixth graders wasstudied by Koval and Pederson (1999). In view of the generallyweak findings from previous research on self-esteem andsmoking, they measured stress in order to look for possible interactions(e.g., perhaps the combination of stress plus lowself-esteem would lead to the most smoking, a pattern consistentwith the predictions of the buffer hypothesis). Self-esteemdid not have any significant relationship to smoking, either byitself or in interaction with stress. This investigation is notablebecause several other variables (including stress, rebelliousness,being in control of other aspects of one’s life, andwhether the child’s mother smoked) did yield significant predictions,and so the failure of self-esteem is not due to anyproblem in measuring smoking.A smaller sample of 183 eighth graders was surveyed byDolcini and Adler (1994). They examined not only self-esteembut also membership in various so-called crowds. Global selfesteemwas not related to cigarette smoking. Similarly, in ayoung adult (college student) sample, Sharp and Getz (1996)found no correlation between self-esteem and smoking.Our study-by-study review suggests the need for a metaanalysisof the literature on self-esteem and smoking. Theremay be a small positive association moderated by gender.Given that the largest study (Abernathy et al., 1995) found alongitudinal effect for females, the possibility that low self-esteemis a risk factor for smoking in girls remains a hypothesis worthyof further study. Still, there was a preponderance of nullfindings, and the few positive findings could conceivably reflectnothing more than self-report bias, or could even indicatethat recognizing oneself as a smoker leads to lower self-esteem.Self-esteem is either irrelevant to the causation of smoking orat best a weak risk factor in girls.Alcohol and Other DrugsAbuse of alcohol and other drugs is one of the most worrisomebehaviors among young people, and there was ample reasonto hope that boosting self-esteem would prevent it (e.g.,California Task Force, 1990). In this view, people turn to drugsfor solace because they feel bad, and low self-esteem is oftenessentially a matter of feeling bad about oneself. Boosting selfesteemwould therefore remove the need for the escapist abuseof these substances.However, other predictions are plausible as well. High selfesteemmight contribute to drug use, too. Gerrard, Gibbons,Reis-Bergan, and Russell (2000) summarized evidence thatyoung people with high self-esteem may become vulnerable todrug and alcohol abuse, albeit for reasons quite different fromthose of their peers with low self-esteem. As we have alreadynoted, people with high self-esteem are prone to a variety of biasesin how they interpret and understand events, and these biasesconspire to make them feel better about themselves. Gerrard30 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.et al. showed that adolescents with high self-esteem systematicallydistort their perceptions of how their parents judge theirdrinking so as to justify their actions. That is, for all adolescents,perceived approval by parents is an important correlateof drinking, but the direction of causation differs as a functionof self-esteem. For teenagers with low self-esteem, perceivedparental approval leads to more drinking. For those with highself-esteem, however, more drinking leads to perceiving greaterparental approval.Gerrard et al. (2000) also showed that individuals with highself-esteem tend to minimize their own vulnerability. They employa variety of cognitive strategies to convince themselvesthat bad things will not or cannot happen to them, and ignoredisagreeable information. Thus, they distance themselves fromthe potentially harmful consequences of risky behavior. Allthese patterns may make people with high self-esteem moreprone to drink, take drugs, and engage in other risks (such asunsafe sex, a point to which we return later in this section). Inshort, these cognitive strategies might at least offset the factors,such as influenceability and need to escape from bad feelings,that could induce people with low self-esteem to do the samethings, and as a result people with low versus high self-esteemmight not differ in their tendency to abuse drugs.The data do not consistently show that adolescent selfesteemcauses or even predicts abuse of alcohol or other drugs.For example, the large-scale study by McGee and Williams(2000) found no relationship between self-esteem (measured atage 9 to 13) and drinking or drug use at age 15. A major longitudinalstudy of heavy drinking in Finland by Poikolainen,Tuulio-Henriksson, Aalto-Setälä, Marttunen, and Lönnqvist(2001) tracked 700 young people from late adolescence (age15–19 at the start of the study) into adulthood 5 years later.They measured both alcohol intake and heavy drinking, withthe latter defined as having 13 or more drinks on one occasion.Self-esteem as measured at the start of the study did not predictgeneral alcohol consumption 5 years later (r .08 for males,r .01 for females), nor did it predict heavy drinking.Likewise, a 7-year longitudinal study of 125 children (aged7 to 13 at the start of the study) by Hill, Shen, Lowers, andLocke (2000) found no effect for self-esteem. In this study, alcoholintake was measured by both self-report and interviewers’assessments. The focus was on whether self-esteem(among other factors) might serve to buffer the individualagainst stresses, and Hill et al. carefully compared familieswith a history of alcoholism against other families without thatrisk factor. Self-esteem did not emerge as a significant predictoror moderator.Even when findings do show a link between alcohol andself-esteem, they are mixed and inconclusive. The large studyby Glendinning (1998; described earlier, in connection withsmoking) found that high self-esteem was associated withmore (rather than less) frequent alcohol consumption amongyoung men but not women, and self-reported episodes of drunkennesswere unrelated to self-esteem. Likewise, in a 4-year studyof 700 students, beginning in middle school, Scheier, Botvin,Griffin, and Diaz (2000) found that students with higher selfesteemin grade 7 reported more rapid rise in alcohol use overthe subsequent years, although this effect appeared only inmore complex analyses; simple analyses yielded no relationshipbetween self-esteem and alcohol use. These authors weresurprised to find that alcohol use was linked to high rather thanlow self-esteem, and they noted that several other patterns intheir data rendered the issue even more obscure. In particular,across the entire group, the several years of middle school weremarked by declining self-esteem and increased drinking, andthese trends were interrelated such that more rapid decreases inself-esteem were accompanied by more severe increases indrinking. Hence, the apparent link between high self-esteemand escalating alcohol abuse might reflect the fact that the studentswith the highest initial self-esteem and the lowest initialdrinking had the largest scope for change. It is a common statisticalillusion for extreme scorers to tend to become less extremeover time, and it often does not indicate any genuinedifference.A cross-sectional survey of college students by Sharp andGetz (1996) found that those who reported having used alcoholin the past month scored higher on self-esteem than those whoreported no alcohol use in the past month. The authors interpretedthis to mean that using alcohol was a social strategy formaking a good impression on peers and therefore in turn resultedin higher self-esteem, but given the correlational natureof their data, the findings could equally well mean that highself-esteem leads to greater willingness to drink alcohol.The study of third and fifth graders by Jackson et al. (1997;described earlier) did find that self-rated self-confidence predictedlesser self-reported alcohol use. With teachers’ ratingsof self-confidence, no effect was found. Moore and Li (1998)found a weak correlation between low self-esteem and illicitdrug use, which remained significant after controlling for othervariables, in a cross-sectional survey of people with disabilities.Alcohol use showed little relation to self-esteem in the study(cited earlier) by Andrews and Duncan (1997). Self-esteemdid, however, predict marijuana use, and in particular wasfound to be a significant mediator between academic motivation(the study’s main focus) and marijuana use. Essentially,the results showed that a decrease in academic motivation ledto a reduction in self-esteem (possibly because poor schoolworkled to social rejection) and a subsequent increase in marijuanause. The relationship between self-esteem and marijuanause was rather weak, however.Wills (1994) challenged earlier findings claiming that lowself-esteem leads to drinking or drug abuse. He proposed thatperceived control over one’s life is more relevant and powerful.Because feeling oneself to be in control of one’s life mightoverlap with both self-esteem and drug use, researchers mightfind correlations between self-esteem and drug use and mistakenlyconclude that self-esteem, rather than feelings of control,VOL. 4, NO. 1, MAY 2003 31

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemwas the important factor. Wills surveyed a large sample ofnearly 1,800 students in eighth grade and then a year later. Hismeasure of perceived control was, in fact, substantially correlatedwith self-esteem (r .67). He then conducted statisticalanalyses to determine which was more powerful at predictinguse of alcohol, tobacco, and marijuana. Because self-esteemand perceived control overlapped with each other, it was necessaryto sort the data statistically in order to establish the uniquecontribution of each (i.e., the effect of each one independent ofthe other). The unique contribution of perceived control wasmore than 6 times as large as the unique contribution of self-esteem.Also, the only significant predictor of increase in substanceabuse over time was lower levels of perceived control.Wills concluded that any findings linking self-esteem to substanceabuse may well be inflated because at bottom they simplyreflect the contribution of perceived control.Interpretation of the findings regarding drinking and drugabuse is probably complicated by the fact that very differentphenomena are lumped together. Some people experiment withdrugs recreationally, whereas others become addicted. Somemay approach the experience out of curiosity or thrill seeking,whereas others may use it to cope with or escape from chronicunhappiness. Emler (2001) concluded that high self-esteemleads to greater willingness to take physical risks, which mayaccount for some of the findings that high self-esteem leads tomore use of alcohol and other drugs.Sexual BehaviorSexual behavior is another sphere that seems to have potentialfor showing links with self-esteem. Some participation insexual behavior, especially at a vulnerable young age, can bethe result of succumbing to influence or even exploitation byothers. High self-esteem might enable people to resist such influence.Such simple predictions do not take account of possiblecomplicating factors, however. For example, a person’s level ofself-esteem may be the result rather than the cause of sexualactivity. Third variables may also play an important role. Forexample, being attractive and popular may lead to both highself-esteem and early or frequent sexual activity. Last, as wenoted earlier, people high in self-esteem tend to discount varioushealth risks (Gerrard et al., 2000), and this can certainly extendto the risks associated with sex. Smith, Gerrard, andGibbons (1997) found that women with high self-esteemtended to dismiss and discount risks of unwanted pregnancymore than other women, and being reminded of the risks onlyincreased the tendency for women with high self-esteem to believethat such misfortunes could not befall them. Similar conclusionswere reached by Burger and Burns (1988), whoshowed that sexually active women with high self-esteem ratedtheir chances of becoming pregnant as lower than those ofother women (including women who were not sexually active!).Burger and Burns concluded that high self-esteem ischaracterized by a pervasive illusion that undesirable consequenceswill not happen to the self.In the California self-esteem project, Crockenberg and Soby(1989) looked for evidence whether self-esteem had any relationshipto teenage sexual activity in general and pregnancy inparticular. They found only one longitudinal and three crosssectionalstudies. The longitudinal study found that high selfesteemled to more sexual activity among males but not females(Jessor & Jessor, 1975). Another study found a weaklink between low self-esteem and sexual activity, but this waslimited to active Mormons (B.C. Miller, Christensen, & Olson,1987). A third study found no relationship at all (Cvetkovich &Grote, 1980), and the last found that high self-esteem was correlatedwith more permissive attitudes toward sex (Herold &Goodwin, 1979). Crockenberg and Soby did, however, thinkthere was some evidence to support the notion that high selfesteemfacilitates use of contraceptives and hence reduces teenpregnancy, although the evidence was mixed. They found onlyone longitudinal study that measured self-esteem and investigatedits relation to subsequent pregnancy, and this study foundno relationship (Vernon, Green, & Frothingham, 1983). Anotherinvestigation, however, used a measure of self-derogationinstead of a standard self-esteem measure, and two differentanalyses (Kaplan, Smith, & Pokorny, 1979; Robbins, Kaplan,& Martin, 1985) found that the girls who derogated themselvesmore during the seventh grade were more likely to have a subsequentpregnancy.The longitudinal study of New Zealand adolescents by Mc-Gee and Williams (2000) found no relationship between selfesteemat age 12 and self-reports of sexual intercourse by theage of 15. Paul, Fitzjohn, Herbison, and Dickson (2000) followedan equally large (N 1,020) sample of New Zealandersfrom age 3 up to age 21. Self-esteem was measured at age 11.At age 21, the participants were asked whether they had hadsexual intercourse before the age of 15. Simple analysesshowed no relationship between self-esteem and early sexualbehavior. More sophisticated (multivariate) analyses still foundno link between self-esteem and early sexual initiation formales. For females, there was a significant effect, although inthe opposite direction from what the advocates of self-esteemmight have hoped: Girls with higher self-esteem at age 11 weremore (rather than less) likely to start having sex by the age of15. This effect was especially remarkable because the otherpredictors of early sexual intercourse tended to be socially undesirableor disadvantageous, such as being in trouble atschool, planning to leave school early, and smoking cigarettesat an early age.Thus, prospective work suggests that self-esteem is irrelevantto becoming sexually active at an early age or, if anything,high self-esteem leads to early sex among girls. In contrast, asurvey of a thousand girls and women by Kalil and Kunz(1999) found that the unmarried teenage mothers were lesslikely than other girls and women to have high self-esteem. Inthis case, self-esteem was measured after the person had en-32 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.gaged in sex and borne a child; in view of the prospective findings,Kalil and Kunz’s results suggest that becoming an unwedteen mother causes a reduction in self-esteem.However, in a large longitudinal survey of young people,Oates (1997) found no relationships between fertility in generaland self-esteem. His sample included more than 10,000young people, aged 14 to 21 at the start of the project and 23 to30 at the end of it, and he sought to test both the idea that low(or high) self-esteem predisposes individuals toward havingchildren and the converse hypothesis that having children affectsself-esteem. No significant relationships were found. Inparticular, there was no relationship between self-esteem andthe number of offspring. Oates acknowledged the theory thatsome people may have children as a way of boosting their selfesteem,but he concluded that even if that is their goal, theseanticipated gains do not seem to materialize.Some work suggests that self-esteem may be relevant tosexual behavior only in selected populations. Berry, Shillington,Peak, and Hohman (2000) used data from a national(United States) longitudinal study of risk factors for adolescentpregnancy. Self-esteem of 5,000 young women was measuredduring early adolescence, and 8 years later the women wereasked whether they had become pregnant before age 19. Highself-esteem reduced the odds of self-reported pregnancy amongBlack and Hispanic females, but not among Whites and AmericanIndians. Even among Blacks and Hispanics, the effect wascharacterized by the authors as slight.At-risk youth were also the focus of Langer and Tubman(1997), who studied a special sample of 120 substance-abusingadolescents. They were specifically interested in whether alcoholand other drug use would predict risky sexual behavior andwhether factors such as self-esteem might moderate the effect—suchas if it was mainly adolescents with low self-esteemwho succumbed to engaging in risky behavior when intoxicated.Self-esteem had no effect.Beyond these findings, a smattering of confusing results hasemerged from different studies. Martin and Knox (1997) foundthat unstable self-esteem, unlike level of self-esteem, was correlatedwith gay and bisexual men’s reports of engaging in unsafeanal sex.People with high self-esteem report more frequent use ofcondoms, which might indicate a greater propensity to practicesafe sex (Hollar & Snizek, 1996). Then again, they also reportgreater frequency of unsafe sex (Hollar & Snizek, 1996). Thereis also some evidence that they may be more likely to haveabortions (Plotnick, 1992). These seemingly contradictory findingsmay be reconciled by proposing that they simply havemore sex in general, a possibility that is supported by otherfindings (Strouse & Buerkel-Rothfuss, 1987). Walsh (1991)found that virginity was linked to low self-esteem among menbut not among women. He also found that self-esteem was positivelycorrelated with number of sex partners; the correlationwas fairly strong for men (r .25) and weak but significant forwomen (r .12). As Walsh noted, the direction of causality remainsunclear, and both directions are plausible, although hefavored the view that high self-esteem makes a person morewilling to initiate sexual encounters because the threat of rejectionis less daunting. Walsh also noted that changes in sexualnorms may produce dramatic changes in the link between selfesteemand sexual activity, depending on whether sexual purityor sexual experience is more highly valued by the society.Homosexuality is generally stigmatized in our society, andone might expect that homosexuals would end up with lowerself-esteem than heterosexuals, although in this case the levelof self-esteem would be the result rather than the cause. Lernerand Galambos (1998) found no difference in self-esteem betweengay and heterosexual adolescents. In contrast, Stokes,Damon, and McKirnan (1997) found that bisexual men withhigh self-esteem were more likely to shift toward a full homosexualidentification than were other bisexual men. This suggestsa causal role for self-esteem, presumably indicating thatsome bisexual men may be resisting homosexuality because ofsocial stigma and other factors, and that the greater self-acceptanceof people high in self-esteem makes them more willing toaccept their homosexuality. When homosexuals are attacked orvictimized for being gay, however, their self-esteem drops afterward(Waldo, Hesson-McInnis, & D’Augelli, 1998). Thus,being victimized leads to lower self-esteem, which is consistentwith the findings on bullying we reported earlier. However,Kendall-Tackett, Williams, and Finkelhor (1993) reviewedmany empirical studies on sexual abuse of children and did notfind evidence that low self-esteem consistently resulted fromsuch abuse (cf. Emler, 2001). They noted that many theoriesabout the impact of sexual abuse on children emphasize damageto self-image or self-esteem, but research findings had notsubstantiated that such damage occurs.Some findings suggest that high self-esteem may reducesexual inhibitions. Herold and Way (1983) found that womenwith high self-esteem were more likely to report that they performedfellatio. Hurlbert and Whittaker (1991) found thatwomen who masturbated had higher self-esteem than otherwomen. Herold, Corbesi, and Collins (1994) found that womenwith high self-esteem were more likely than other women to gotopless on Australian beaches. Although these findings are correlational,we regard it as a priori unlikely that high self-esteemis the result of performing fellatio, masturbating, and showingoff one’s breasts in public, and so a causal role for self-esteemseems more plausible. Conversely, low self-esteem is correlatedwith erectile dysfunction among men (Herer & Holzapfel,1993), but insofar as most erectile dysfunction is nowconsidered to have biological causes, self-esteem seems morelikely to be the result than the cause in this case.All in all, the results do not support the simple view that lowself-esteem predisposes people to more or earlier sexual activity.If anything, people with high self-esteem are less inhibited,more willing to disregard risks, and more prone to engage insex. Many of these findings could stem from the greater popularityof people with high self-esteem (popularity could causeVOL. 4, NO. 1, MAY 2003 33

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteemboth high self-esteem and more sex). They may also reflectgreater initiative among people high in self-esteem. Not surprisingly,however, bad sexual experiences and unwanted pregnanciesmay well lower self-esteem.Eating DisordersThe literature on eating disorders very consistently documentsthat individuals with eating disorders typically have verylow levels of self-esteem. A recent and particularly impressivedemonstration of this association was reported by French et al.(2001), who examined the correlates of binge-purge andweight-loss behaviors among a sample of more than 48,000girls and 47,000 boys in grades 6 through 12 who lived in 213cities and towns in the United States. For both genders, selfesteemand similar variables were the strongest predictors of eatingdisorders. Similar results have been obtained in many otherstudies, including one by Williams et al. (1993), who found thatanorexic and bulimic patients have lower levels of self-esteemthan obese dieters, nonobese dieters, and normal controls.Anorexics’ thinking is pervaded by a discrepancy betweentheir own view of themselves and other people’s view of them:They frequently have many accomplishments, abilities, and interests,but according to anorexia theorists, they are unaware oftheir successes and capabilities. Bruch (1962, 1975, 1978)identified in these patients a paralyzing underlying sense of ineffectivenessthat pervades their thinking and activities. Bersand Quinlan (1992) showed empirically that this is true, stating,“This disparity between interests and perceived abilitiessupports the common clinical observation that anorexics generallyfeel incompetent and unable to perform well, althoughthey pursue many activities and objectively could claim manyaccomplishments” (p. 428).Perhaps because bulimia is a much more pervasive problemthan anorexia, more research has been done to investigate possiblecauses of bulimia. Mintz and Betz (1988) investigatedself-esteem differences in a sample of undergraduate womenwho were categorized in one of the following categories: normals,bulimics, bingers (women who eat large amounts of foodbut do not engage in compensatory behaviors), purgers (thosewho purge but do not really eat huge amounts of food), chronicdieters, and subthreshold bulimics. They found that bulimicshad much lower self-esteem than any of the other groups. Thus,the impact of self-esteem may be tied specifically to the patternof bingeing and purging.Bulik, Wade, and Kendler (2000) studied the relation betweenself-esteem and bulimia in monozygotic (“identical”)twins. Although monozygotic twins have the same genetic predispositionto bulimia, it is sometimes the case that one twinhas the disorder and the other does not. In a sample of suchtwin pairs, Bulik et al. found that the twin with bulimic symptomshad significantly lower self-esteem than the other. In alaboratory study of chronic dieters who were given a high-caloriemilkshake to drink, those with low (but not high) selfesteemate considerably more ice cream later (Polivy, Heatherton,& Herman, 1988). However consistent these findings are, arecent study calls into question whether raising self-esteemmight be useful for treating already-established eating disorderpatterns. In this study (Safer, Lively, Telch, & Agras, 2002),women who were binge eaters (though not purgers) receivedbehavior therapy for 20 weeks and then 6 months later were reassessed.At the follow-up, the difference in self-esteem scoresbetween women who relapsed and those who did not was nonsignificant,and, indeed, those who relapsed had slightly higherself-esteem.Although researchers have found fairly consistent associationsbetween low self-esteem and eating pathology, there issome debate as to whether low self-esteem is a cause or consequenceof disordered eating. There are some prospective studiesthat suggest self-esteem may play a causal role. For example,Button, Sonuga-Barke, Davies, and Thompson (1996) assessedself-esteem in 400 schoolgirls aged 11 to 12 and found thatthose with lower levels of self-esteem were significantly morelikely to have developed an eating disorder at age 15 to 16.There is also some evidence that self-esteem may affect prognosis;in a 4-year prospective study of bulimic patients, lowlevels of self-esteem at admission were predictive of pooreroutcomes (van-der-Ham, van-Strein, & van-Engeland, 1998).Heatherton and Polivy’s (1992) spiral model of binge eating,however, posits a bidirectional relationship between dietingand self-esteem. They stated thatbecause each dietary failure may produce lower self-esteem and becauselower self-esteem may, in turn, make dietary failure more likely,individuals who undertake chronic dieting may enter a spiral in whicheach failure at dieting produces greater negative affect and precludeseither successful acceptance or successful alteration of their bodies.(p. 139)Another factor that must be considered is that women whobecome bulimic have additional problems that contribute totheir vulnerability, and to the extent that these other problemsare correlated with low self-esteem, the importance of selfesteemper se may be exaggerated when they are not taken intoaccount. The best way to examine this hypothesis is to assessthe relations between several vulnerability factors (includingself-esteem) and the continuum of bulimic symptoms amongwomen who have not sought treatment for bulimia. If this canbe done in a prospective study, the causal direction of any linkbetween low self-esteem and bulimic symptoms can be tested.This method was employed by Vohs and her colleagues(Vohs, Bardone, Joiner, Abramson, & Heatherton, 1999; Vohset al., 2001; for a summary, see Vohs et al., 2002), who conductedtwo prospective studies with a sample totaling morethan 400. Vohs et al. (1999) measured self-esteem, perfectionism,feeling overweight, and bulimic symptoms at the initialassessment and then measured bulimic symptoms again an averageof 9 months later. They found that although self-esteem34 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.scores were significantly correlated with bulimic symptoms(.52 and .36 at the first and second assessments, respectively),self-esteem was not a significant predictor of change inbulimic symptoms in a statistical model that also included theeffects of perfectionism and feeling overweight. Bulimia wasnot predicted by the combination of level of self-esteem withlevel of perfectionism or with feeling overweight. Analysesshowed, however, that a combination of all three factors was asignificant predictor of increases in bulimic symptoms. Specifically,the combination of high perfectionism and feeling overweightwas predictive of increased bulimic symptoms across 9months, but only among low self-esteem women. This findingsuggests that low self-esteem is a risk factor for bulimic symptomatology,but only when it is accompanied by exceedinglyhigh standards for oneself (perfectionism) and a feeling thatthose standards are not being met (feeling overweight).A second study, across a 5-week period, showed the samepattern of results (Vohs et al., 2001). In this study, there weresimple associations between self-esteem scores and bulimicsymptoms, but the only statistical factor to predict increased bulimicsymptoms was the combination of having low self-esteem,having high body dissatisfaction, and being highly perfectionistic.This study also examined the development of anxiety anddepressive symptoms, and analyses showed that self-esteem didnot, on its own, predict change in depression or anxiety whenperfectionism and body dissatisfaction were taken into account.However, the results for depression mirrored the pattern seenfor bulimia: The greatest increase in depressive symptoms occurredamong women who had high perfectionism scores, highbody dissatisfaction, and low self-esteem.Taken together, these studies suggest that self-esteem doesplay a role in various eating problems. However, it is also importantto consider that the role of self-esteem may be muchmore complex than is suggested by the simple correlations betweenconstructs. Bulimics, in particular, may have low selfesteem,although low self-esteem may be both a predisposingcause and a consequence of the disordered eating. A recentstudy of obese women supports this idea. Matz, Foster, Faith,and Wadden (2002) found that self-esteem was second only tobeing teased in the ability to predict degree of body-image dissatisfaction.However, we again note that the influence of selfesteemon pathological eating patterns may be a consequenceof the presence of other risk factors (such as perfectionism andbody dissatisfaction) or of the interaction between low selfesteemand these other factors.ConclusionMost studies reviewed in this section relied on self-report,and the possibility of response bias is therefore substantial.Even so, the studies do not provide a great deal of evidence thathigh self-esteem can prevent undesirable outcomes. The mostpromising possibility is that high self-esteem might prevent bulimia,and there are some links to longevity and physical healththat seem well worth further study.Most studies on self-esteem and smoking have failed to findany significant relationship, even with very large samples andthe correspondingly high statistical power. The studies thathave found a relationship have tended to find it only in someanalyses, such as only among girls and not boys, or only forsome ages, or only for current smokers and not for ex-smokersor experimental smokers.Self-esteem also does not seem relevant to use of alcoholand other drugs, although we did not find much highly rigorousdata on use of illegal drugs, and it is conceivable that furtherstudies might paint a different picture. Large, longitudinal investigationshave tended to yield no relationship between selfesteemand either drinking in general or heavy, problem drinkingin particular. The few studies that have found links point indifferent directions, with some linking alcohol use to higherself-esteem and others to lower self-esteem. Moreover, even ifsome correlations between low self-esteem and substanceabuse could be found, they might well turn out to be based onthird variables, such as perceived control (Wills, 1994). Whateverthe causes of alcohol abuse and drug addiction, low selfesteemper se does not appear to be one of them.Self-esteem does not appear to prevent early sexual activityor teen pregnancy. Some studies have found self-esteem to beunrelated to sexuality. Others have yielded small effects thatsometimes point in contrary directions. One promising patternsuggests that high self-esteem reduces sexual inhibitions, enablingwomen to engage in various sexual practices morefreely and enabling people to accept their homosexual tendencies.Still, the causal inference is speculative, and it is quiteplausible that greater sexual freedom leads to popularity, whichin turn boosts self-esteem, as the sociometer model might predict.Indeed, there are various signs supporting the hypothesisthat popularity is the hidden variable behind the correlationsbetween high self-esteem and greater sexual activity.The data do, however, suggest a role for self-esteem with respectto disordered eating patterns. Concurrent and prospectivefindings indicate that low self-esteem is a risk factor in disorderedeating, with evidence suggesting that the development ofbulimic symptoms may be affected both directly by the presenceof low self-esteem and indirectly by the interaction of lowself-esteem with other factors (e.g., Vohs et al., 1999, 2001).However, eating disorders are usually preceded by chronic dietingand body dissatisfaction, which themselves are related tolow self-esteem (see, e.g., Heatherton & Polivy, 1992; Matz etal., 2002). And it is also important to note that some womenwith eating disorders also have other disorders, such as anxietyand depression, that are also related to low self-esteem. Inshort, there is a complex set of relationships through whichself-esteem has its effects on vulnerability to eating disorders.Nonetheless, on the whole, we conclude that low self-esteem isa concurrent and prospective risk factor for eating disordersymptoms.VOL. 4, NO. 1, MAY 2003 35

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-EsteemGENERAL DISCUSSION AND CONCLUSIONSOur goal was to survey the research literature in order to assesswhether self-esteem has important consequences. The taskwas complicated by four main factors: (a) the many thousandsof articles alluding to self-esteem; (b) the tendency of peoplehigh in self-esteem to rate themselves as superior on many dimensions(and the floccinaucinihilipilification among peoplewith low self-esteem); (c) the difficulty of establishing the directionof causal relationships; and (d) the heterogeneity ofhigh self-esteem. When possible, we restricted our search tostudies with objective measures, and we have emphasized largelongitudinal studies that offered some opportunity to assess thedirection of causality.Our particular interest was in the possible benefits of highself-esteem and the corresponding costs of low self-esteem. Aswe note later, this is not the only possible theoretical perspective,but it is the one that has characterized the self-esteemmovement and indeed our own initial interest in self-esteem.Does high self-esteem make life better? Should parents, teachers,or anybody else seek to boost self-esteem whenever possible?We begin with a summary of our main findings.Summary of FindingsHigh self-esteem does not reliably cause any improvementin academic performance. If anything, high self-esteem may bethe result (but only weakly) of doing well in school. Other factorsmay underlie both self-esteem and academic performance.People with high self-esteem do better than other people onsome jobs and tasks, although most laboratory studies and manyfield studies have found no difference. Quite possibly, occupationalsuccess leads to high self-esteem rather than the reverse.High self-esteem does improve persistence in the face offailure, especially when persistence is an adaptive strategy.People with high self-esteem are more willing than others tochoose their own strategies, and they are more responsive tosituational cues indicating when to persist and when to moveon to a more promising alternative.People high in self-esteem regard themselves as better likedand more popular than others, but most of these advantages existmainly in their own minds, and objective data (such as ratingsby peers) generally fail to confirm them. In some cases,such as after an ego threat, people with high self-esteem are actuallydisliked more than others.There is relatively little known about the impact of self-esteemon close relationships. The evidence that is available indicatesboth advantages and disadvantages. Thus, wives complain moreabout husbands with low than high self-esteem, although the directionof causality is unclear. People with high self-esteem aremore willing than others to terminate a relationship, but there isno sign that their relationships actually suffer more or endsooner.People with high self-esteem sometimes perform better thanpeople with low self-esteem in groups. They speak up moreand are recognized by peers as contributing more. There aresome weak simple correlations between self-esteem and leadership,but analyses that control for other factors have foundthat self-esteem has little in the way of direct and unique causalimpact. The possibility remains, however, that self-esteem hasindirect effects that are mediated by other factors, such as leadershipefficacy.People with high self-esteem have a stronger tendency thanthose with low self-esteem to judge and treat their own groupsmore favorably than out-groups. (This may be regarded eitherpositively, as in supporting group pride, or negatively, as incontributing to prejudice and discrimination.)Self-esteem is essentially unrelated to aggression, althoughthis overall pattern may conceal divergent trends for differentkinds of high self-esteem. Some categories of high self-esteem(such as defensiveness or narcissism) are associated with heightenedaggression. High self-esteem predicts being a bully and supportinga bully–but a different category of high self-esteempredicts defending victims against bullies. High self-esteem maythus amplify both prosocial and antisocial tendencies. Low selfesteemis associated with victimization, although whether thelow self-esteem is the cause or consequence is not known.Self-esteem has at best a weak relationship to delinquency.Many findings are tainted by possible self-report bias and byinconsistencies across studies. Some studies suggest that lowself-esteem predisposes people to engage in more delinquentbehavior later on (at least, by their own report), whereas otherstudies have found no effect. However, one recent study providedgood evidence that low self-esteem leads to delinquency(Trzesniewski et al., 2002).Likewise, there is little to connect self-esteem directly to antisocialbehavior. If anything, some subcategories of high selfesteemare associated with antisocial behavior. Again, self-esteemmay magnify both prosocial and antisocial tendencies.Happiness appears to be the most desirable correlate of highself-esteem. Although research is needed to establish causalityand to control for other variables, it seems quite possible thathigh self-esteem contributes to making people happy. Low selfesteemis linked to depression and may be a risk factor for it.However, the relationship is weak, inconsistent, and conditionalon other variables.The buffer hypothesis proposes that high self-esteem helpspeople cope with stress and adversity. There are some positivefindings to support this view. Other findings, however, indicatethat self-esteem is more relevant under low than high stress(low self-esteem poisons the good times). Yet other findings indicateno effect of self-esteem in either direction. Still, no findingsshow worse outcomes or poorer coping among peoplewith high self-esteem than among those with low self-esteem.Thus, self-esteem may contribute to coping and adjustment afterstress or trauma, although the precise nature of the relationshipmay be complicated and may depend on other factors.Self-esteem is largely irrelevant to smoking. At best, to asmall degree high self-esteem may lead to less smoking among36 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.young females. Given the null findings in some large studies,self-esteem is probably not worth considering as a cause ofsmoking. Use of alcohol or other drugs also does not show aconsistent relationship to self-esteem. If anything, young peoplewith high self-esteem may be more willing than others toexperiment.High self-esteem does not prevent early, extensive, or riskysexual activity. If anything, people with high self-esteem displayfewer inhibitions, more disregard for risks, and greatersexual initiative.Low self-esteem is a risk factor that can contribute to eatingdisorders, especially bulimia. Its relation to bulimia may be especiallystrong in the presence of other factors, such as bodydissatisfaction and perfectionism.We conclude our summary of findings with some generalpoints.• With the exception of the link to happiness, most of the effectsare weak to modest. Self-esteem is thus not a majorpredictor or cause of almost anything (again, with the possibleexception of happiness).• Moreover, the effects of self-esteem become weaker as thecriteria for evidence become more objective. It is perhaps noaccident that the strongest apparent benefit of self-esteemhas been found for the most subjective outcome, namely,happiness. As we noted at the outset, people with high selfesteemseem sincerely to believe they are smarter, more accomplished,more popular and likable, and more attractivethan other people, but some of those apparent advantages areillusory. But happiness can coexist well with illusions andmay even be supported by them, so the fact that objectivedata disconfirm many of the subjective advantages peoplewith high self-esteem believe they have is not necessarily areason to dismiss self-esteem as useless.• The research repeatedly attests to the heterogeneity of highself-esteem, and many researchers have invoked some sort ofdistinction between being conceited, narcissistic, and defensive,on the one hand, as opposed to accepting oneself withan accurate appreciation of one’s strengths and worth, on theother.• The effects of self-esteem are often enmeshed with the effectsof other correlated variables, and so some apparent effectsof self-esteem vanish when other variables are controlled for.The appropriateness of controlling for these other variablesis debatable, however. One could argue that self-esteem deservescredit for so-called indirect effects, even if other variablesare more directly related to the outcome.• A final point, which is related to the others, is that it is farfrom clear that interventions aimed at boosting self-esteemwill be sufficient to produce positive outcomes. The seemingbenefits of self-esteem could be a product of other factorsthat overlap with self-esteem, of subjective bias, and of reciprocalcausal relationships.Implications for Understanding the Natureof Self-EsteemThe benefits of high self-esteem can be tentatively summarizedin terms of two main themes, although these remainsomewhat speculative and are contingent on further work supportingthe conclusions we just summarized. First, high selfesteemappears to operate as a stock of positive feelings thatcan be a valuable resource under some conditions. In the faceof failure or stress, people with high self-esteem seem able tobounce back better than people with low self-esteem. The generalpattern of being happier and less depressed indicates areadiness to feel good. People with low self-esteem lack thisstock of good feelings and as a result are more vulnerable.Second, high self-esteem appears linked to greater initiative.We suggested that people with high self-esteem are more proneto both prosocial and antisocial actions (e.g., both bullying anddefending victims against bullies), compared with people withlow self-esteem. They initiate interactions and relationships(and perhaps exit them, too). They speak up in groups. Theyexperiment with sex and perhaps drugs. They try harder in responseto initial failure, but they are also willing to switch to anew line of endeavor if the present one seems unpromising.There appear to be relatively few personal costs to high selfesteem.If anything, the costs of high self-esteem (and especiallyof certain subcategories of high self-esteem, such as narcissism)are borne by other people. People high in self-esteemor narcissism are prone to bully others, to retaliate aggressively,and to be prejudiced against out-group members. Selfenhancersare sometimes annoying or obnoxious to others.They may be willing to cheat and perform other antisocial,self-serving acts. The most palpable cost to the self is that ofoverconfident risk taking (see Emler, 2001), and the evidencesupporting this possibility is quite sparse.Thus, a tentative formula to integrate the diverse findings isthat self-esteem confers some benefits on the self, including feelingquite good, while its costs accrue to others. Having a firmsense of privileged superiority over everyone else may well be apleasant, rewarding state, but having to live or work with someonewho holds such an inflated self-view may have its drawbacks.Indeed, the socially disruptive consequences of egotism may explainwhy people tend to be modest and self-effacing when interactingwith friends (Tice, Butler, Muraven, & Stillwell, 1995) orwhen living in cultures characterized by stable, interdependent relationships(Heine et al., 1999). If self-esteem does benefit theself while carrying some cost to others, then its pursuit is moresuited to an individualistic than a collectivistic culture.What About Accurate (Not High) Self-Esteem?In our review of the literature, we focused on investigatingthe benefits of high self-esteem. Given that these benefits arelimited at best, it may be appropriate to rethink the basic question.Perhaps it is more valuable and adaptive to understandoneself honestly and accurately, even when this means feelingVOL. 4, NO. 1, MAY 2003 37

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-Esteembad about oneself when that is warranted by unethical, harmful,socially undesirable, or otherwise inappropriate behavior.According to this view, self-esteem could serve valuable andhelpful functions (such as for managing one’s life) insofar as itis based on an accurate, rather than inflated, assessment ofone’s characteristics. That would mean that accurate selfknowledgewould be more useful than high self-esteem.Such an argument has been suggested by Kirkpatrick andEllis (2001). They presented their ideas within an evolutionaryframework, although other frameworks might also work. Theirpoint was that an accurate understanding of self can save considerabletime and energy. In the pursuit of mates, for example,it could spare people the wasted effort (not to mention disappointmentand heartbreak) of pursuing mates who will rejectthem. By the same token, accurate self-knowledge might helppeople know what courses to take in school or what occupationsto pursue, avoiding both ones that are too difficult (whichwould produce failure) and those that would be too easy (andhence would be unrewarding). Although people might prefer tohold highly favorable views of themselves, accurate viewswould almost certainly be more useful, insofar as accurate informationis conducive to more effective decision making.A new emphasis on self-esteem that accurately reflects capabilitiesand interpersonal characteristics would hardly meshwith the goals and efforts of the self-esteem movement. Itwould require that interventions be aimed at lowering somepeople’s self-esteem, even while boosting others’. In fact,given the evidence cited earlier that self-esteem is mildly inflatedin the modern American population, the pursuit of accurateself-esteem might well entail lowering self-esteem morethan raising it. Conceivably, however, a new emphasis on cultivatingaccurate self-understanding would pay off in terms ofbetter choices. Then again, if high self-esteem produces happiness,the relative value of happiness versus better choices basedon accurate information could also be debated. As scientists,we are inclined to favor the pursuit of truth above all else, butwe can recognize that some people might prefer self-flatteringillusions over accurate knowledge.Implications for PolicyUltimately, should our society try to boost people’s self-esteem?A case can be made for either answer. On the positive side,high self-esteem does have more benefits than costs, even ifboth are limited. If it produces happiness, initiative, and resilience,those are certainly positive outcomes that any parentmight wish for his or her child. On the negative side, it is notclear that simply boosting self-esteem will in turn produce alasting increase in happiness, initiative, resilience, or any otherdesirable outcome. There is even the possibility that boostingself-esteem will end up fostering narcissism or one of the othersocially undesirable forms of high self-esteem.In our view, a crucial issue for both research and policy isthe heterogeneity of high self-esteem. Only a few of the manystudies we reviewed distinguished carefully between differentcategories of favorable self-regard, yet these few often foundthe distinction to be quite powerful. Indeed, sometimes oppositerelationships existed, which summed to a deceptive appearanceof no correlation when all forms of high self-esteemwere lumped together. We recommend that researchers interestedin self-esteem begin paying closer attention to narcissism,self-deception, stability of self-esteem, and other distinctionswithin the broad category of self-esteem. For practitioners, theimplication is that high self-esteem as such may not be worthcultivating, because some forms of it are unhelpful and possiblyharmful—but other forms or versions of it may be quitebeneficial.Although splitting high self-esteem into categories may beseen as fatal to the self-esteem movement, a more optimisticand charitable conclusion is that such splitting may yet hold thekey to reaping some of the benefits that self-esteem advocateshave long promised. Self-esteem per se is not the social panaceathat many people once hoped it was, but if researchers caneffectively identify the adaptive, desirable subcategories ofhigh self-esteem, it may yet be possible to say that certain waysof thinking about oneself can cause good things to happen. Topropose how thinking of oneself can bring about good outcomesis to go beyond the data we have reviewed, and so weconfine such reflections to a final, overtly speculative section,which is presented next.Concluding ReflectionsOur views on the merits of boosting self-esteem have gonethrough multiple changes, and were further revised during theprocess of compiling this review. We conclude by offering ourcurrent view, with frank acknowledgment that it should be regardedmore as informed expert opinion than as an unassailablesummary of proven facts.In some ways, the grandfather of the self-esteem movementwas Carl Rogers, who promoted the idea of “unconditionalpositive regard” as a way of helping children avoid the feelingthat their parents might stop loving them if they failed to performup to high standards. Sadly, over time unconditional positiveregard has taken the form of suggesting that parents andteachers should never criticize children and indeed shouldpraise children even for mediocre or trivial accomplishments,or just for being themselves. Always praising and never criticizingmay feel good to everyone concerned, but the data wehave reviewed do not show that such an approach will producedesirable outcomes.We have already indicated what the data do show. They suggestthat the benefits of high self-esteem are far fewer andweaker than proponents of self-esteem had hoped. Still, thereare some benefits, and the costs to the individual do not outweighthem. The possible costs to society, such as from havingsome people regard themselves as superior to others and henceentitled to exploit their fellows or demand preferential treat-38 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.ment, may be another matter. Even so, these costs are associatedwith only particular subcategories of high self-esteem.The heterogeneity of high self-esteem is central to ourthinking, and it suggests that self-esteem per se is the wrongfocus. There are many ways to think well of oneself, and someof these produce more desirable outcomes than others. Evensuch leaders of the self-esteem movement as Nathaniel Brandenhave begun to speak of the need for self-esteem to be linked inparticular ways to other aspects of life (such as moral virtue orlegitimate achievement). To them, perhaps, this strategy is amatter of cultivating genuine self-esteem instead of other forms.To us, all favorable views of self may be genuine self-esteem,and so self-esteem is inherently too broad a focus.Thus, we recommend that instead of trying to find the right“kind” of self-esteem and perhaps dismissing other kinds asunreal, policymakers and practitioners should seek the right usageof self-esteem. High self-esteem feels good and fosters initiative.It may still prove a useful tool to promote success andvirtue, but it should be clearly and explicitly linked to desirablebehavior. After all, Hitler had very high self-esteem and plentyof initiative, too, but those were hardly guarantees of ethicalbehavior. He attracted followers by offering them self-esteemthat was not tied to achievement or ethical behavior—rather, hetold them that they were superior beings simply by virtue ofbeing themselves, members of the so-called Master Race, anidea that undoubtedly had a broad, seductive appeal. We havefound no data to indicate that indiscriminately promoting selfesteemin today’s children or adults, just for being themselves,has any benefits beyond that seductive pleasure.Hence, we think self-esteem should be used in a limited wayas one of a cluster of factors to promote positive outcomes. Itshould not be an end in itself. Raising self-esteem will not byitself make young people perform better in school, obey thelaw, stay out of trouble, get along better with their fellows, orrespect the rights of others, among many other desirable outcomes.However, it does seem appropriate to try to boost people’sself-esteem as a reward for ethical behavior and worthyachievements. Although that may sound banal, we think it willrequire a basic change in many self-esteem programs, whichnow seek to boost everyone’s self-esteem without demandingappropriate behavior first.Using self-esteem as a reward rather than an entitlementseems most appropriate to us. To be sure, there may still be aplace for unconditional positive regard, such as when a parentshows love for a child independent of achievement. But whenachievement or virtue is involved, self-esteem should be conditionalupon it. A favorable view of self should be promoted onthe basis of performing well and behaving morally. By thesame token, we think it appropriate and even essential to criticizeharmful or unethical behavior and lazy or deficient performance,without worrying that someone’s self-esteem might bereduced.In particular, we think that success in modern society dependson lifelong learning and improvement—academically,socially, culturally, and occupationally. We encourage linkingself-esteem to learning and improvement. Learning is most effectivewhen one receives both praise and criticism, contingenton current performance. The praise-only regimen of the selfesteemmovement is ultimately no more effective for learningthan the criticism-only regimen of the previous era (althoughpraise-only may feel much more pleasant for all concerned).Praise that bolsters self-esteem in recognition of good performancecan be a useful tool to facilitate learning and further improveperformance in the future. Praising all the children justfor being themselves, in contrast, simply devalues praise andconfuses the young people as to what the legitimate standardsare. In the long run, if such indiscriminate praise has any effecton self-esteem, it seems more likely to contribute to narcissismor other forms of inflated self-esteem than to the kind of selfesteemthat will be best for the individual and for society.A focus on improvement, in particular, allows people tocompare themselves against themselves so that they do nothave to boost themselves at the expense of others. Improvementstrikes us as the ideal condition for boosting self-esteem:As the person performs or behaves better, self-esteem is encouragedto rise, and the net effect will be to reinforce bothgood behavior and improvement. Those outcomes are conduciveto both the happiness of the individual and the bettermentof society.Acknowledgments—We gratefully acknowledge financial support fromthe National Institutes of Health (Grants MH 57039 and MH 12794), theFoundations’ Fund for Research in Psychiatry, and the Social Sciences andHumanities Research Council of Canada. We are indebted to Karyn Cirinofor preparing the reference list.REFERENCESAbernathy, T.J., Massad, L., & Romano-Dwyer, L. (1995). The relationship betweensmoking and self-esteem. Adolescence, 30, 899–907.Aberson, C.L., Healy, M., & Romero, V. (2000). Ingroup bias and self-esteem:A meta-analysis. Personality and Social Psychology Review, 4, 157–173.Abramson, L.Y., Metalsky, G.I., & Alloy, L.B. (1989). Hopelessness depression:A theory-based subtype of depression. Psychological Review, 96,358–372.Adams, G.R., Ryan, B.A., Ketsetzis, M., & Keating L. (2000). Rule complianceand peer sociability: A study of family process, school-forced parent-childinteractions, and children’s classroom behavior. Journal ofFamily Psychology, 14, 237–250.American Psychiatric Association. (2000). Diagnostic and statistical manualof mental disorders (4th ed., text revision). Washington, DC: Author.Andrews, J.A., & Duncan, S.C. (1997). Examining the reciprocal relation betweenacademic motivation and substance use: Effects of family relationships,self-esteem, and general deviance. Journal of Behavioral Medicine,20, 523–549.Arndt, J., & Goldenberg, J.L. (2002). From threat to sweat: The role of physiologicalarousal in the motivation to maintain self-esteem. In A. Tesser,D.A. Stapel, & J.V. Wood (Eds.), Self and motivation: Emerging psychologicalperspectives (pp. 43–69). Washington, DC: American PsychologicalAssociation.Bachman, J.G., & O’Malley, P.M. (1977). Self-esteem in young men: A longitudinalanalysis of the impact of educational and occupational attainment.Journal of Personality and Social Psychology, 35, 365–380.VOL. 4, NO. 1, MAY 2003 39

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-EsteemBachman, J.G., & O’Malley, P.M. (1986). Self-concepts, self-esteem, and educationalexperiences: The frog pond revisited (again). Journal of Personalityand Social Psychology, 50, 35–46.Bandura, A. (1989). Self-regulation of motivation and action through internalstandards and goal systems. In L. Pervin (Ed.), Goal concepts in personalityand social psychology (pp. 19–85). Hillsdale, NJ: Erlbaum.Battistich, V., Solomon, D., & Delucchi, K. (1993). Interaction processes andstudent outcomes in cooperative learning groups. The Elementary SchoolJournal, 94, 19–32.Baumeister, R.F. (1987). How the self became a problem: A psychological reviewof historical research. Journal of Personality and Social Psychology,52, 163–176.Baumeister, R.F. (1989). The optimal margin of illusion. Journal of Social andClinical Psychology, 8, 176–189.Baumeister, R.F. (1993). Self-esteem: The puzzle of low self-regard. New York:Plenum Press.Baumeister, R.F., Dale, K., & Sommer, K.L. (1998). Freudian defense mechanismsand empirical findings in modern social psychology: Reaction formation,projection, displacement, undoing, isolation, sublimation, anddenial. Journal of Personality, 66, 1081–1124.Baumeister, R.F., Heatherton, T.F., & Tice, D.M. (1993). When ego threats leadto self-regulation failure: Negative consequences of high self-esteem.Journal of Personality and Social Psychology, 64, 141–156.Baumeister, R.F., Smart, L., & Boden, J.M. (1996). Relation of threatened egotismto violence and aggression: The dark side of high self-esteem. PsychologicalReview, 103, 5–33.Baumeister, R.F., Tice, D.M., & Hutton, D.G. (1989). Self-presentational motivationsand personality differences in self-esteem. Journal of Personality,57, 547–579.Berry, E.H., Shillington, A.M., Peak, T., & Hohman, M.M. (2000). Multi-ethniccomparison of risk and protective factors for adolescent pregnancy.Child and Adolescent Social Work Journal, 17, 79–96.Bers, S.A., & Quinlan, D.M. (1992). Perceived-competence deficit in anorexianervosa. Journal of Abnormal Psychology, 101, 423–431.Bishop, J.A., & Inderbitzen, H.M. (1995). Peer acceptance and friendship: Aninvestigation of their relation to self-esteem. Journal of Early Adolescence,15, 476–489.Blascovich, J., & Tomaka, J. (1991). Measures of self-esteem. In J. Robinson,P. Shaver, & L. Wrightsman (Eds.), Measures of personality and socialpsychological attitudes (pp. 115–160). San Diego, CA: Academic Press.Bonanno, G.A., Field, N.P., Kovacevic, A., & Kaltman, S. (2002). Self-enhancementas a buffer against extreme adversity: Civil war in Bosnia andtraumatic loss in the United States. Personality and Social PsychologyBulletin, 28, 184–196.Bowles, T. (1999). Focusing on time orientation to explain adolescent self conceptand academic achievement: Part II. Testing a model. Journal of AppliedHealth Behaviour, 1, 1–8.Branden, N. (1984, August–September). In defense of self. Association for HumanisticPsychology, pp. 12–13.Branden, N. (1994). The six pillars of self-esteem. New York: Bantam Books.Brewer, M.B., & Kramer, R.M. (1985). The psychology of intergroup attitudesand behavior. Annual Review of Psychology, 36, 219–243.Brockner, J. (1983). Low self-esteem and behavioral plasticity: Some implications.In L. Wheeler & P. Shaver (Eds.), Review of personality and socialpsychology, Vol. 4 (pp. 237–271). Beverly Hills, CA: Sage.Brockner, J., & Lloyd, K. (1986). Self-esteem and likability: Separating factfrom fantasy. Journal of Research in Personality, 20, 496–508.Brown, J.D. (1986). Evaluations of self and others: Self-enhancement biases insocial judgment. Social Cognition, 4, 353–376.Brown, J.D. (1998). The self. Boston: McGraw-Hill.Bruch, H. (1962). Perceptual and conceptual disturbance in anorexia nervosa.Psychosomatic Medicine, 24, 187–194.Bruch, H. (1975). Obesity and anorexia nervosa: Psychosocial aspects. Australianand New Zealand Journal of Psychiatry, 9, 159–161.Bruch, H. (1978). The golden cage: The enigma of anorexia nervosa. Cambridge,MA: Harvard University Press.Buhrmester, D., Furman, W., Wittenberg, M.T., & Reis, H.T. (1988). Five domainsof interpersonal competence in peer relationships. Journal of Personalityand Social Psychology, 55, 991–1008.Bulik, C.M., Wade, T.D., & Kendler, K.S. (2000). Characteristics of monozygotictwins discordant for bulimia nervosa. International Journal of EatingDisorders, 29, 1–10.Burger, J.M., & Burns, L. (1988). The illusion of unique invulnerability andthe use of effective contraception. Personality and Social PsychologyBulletin, 14, 264–270.Bushman, B.J., & Baumeister, R.F. (1998). Threatened egotism, narcissism,self-esteem, and direct and displaced aggression: Does self-love or selfhatelead to violence? Journal of Personality and Social Psychology, 75,219–229.Button, E.J., Sonuga-Barke, E.J.S., Davies, J., & Thompson, M. (1996). A prospectivestudy of self-esteem in the prediction of eating problems in adolescentschoolgirls: Questionnaire findings. British Journal of ClinicalPsychology, 35, 193–203.Cadinu, M.R., & Rothbart, M. (1996). Self-anchoring and differentiation processin the minimal group setting. Journal of Personality and Social Psychology,70, 661–677.California Task Force to Promote Self-Esteem and Personal and Social Responsibility.(1990). Toward a state of self-esteem. Sacramento: CaliforniaState Department of Education.Campbell, J.D., & Fairey, P. (1985). Effects of self-esteem, hypothetical explanations,and verbalization of expectancies on future performance. Journalof Personality and Social Psychology, 48, 1097–1111.Campbell, J.D., & Fehr, B.A. (1990). Self-esteem and perceptions of conveyedimpressions: Is negative affectivity associated with greater realism? Journalof Personality and Social Psychology, 58, 122–133.Campbell, W.K., Foster, C.A., & Finkel, E.J. (2002). Does self-love lead tolove for others? A story of narcissistic game playing. Journal of Personalityand Social Psychology, 83, 340–354.Chemers, M.M., Watson, C.B., & May, S.T. (2000). Dispositional affect andleadership effectiveness: A comparison of self-esteem, optimism, and efficacy.Personality and Social Psychology Bulletin, 26, 267–277.Clement, R.W., & Krueger, J. (2002). Social categorization moderates socialprojection. Journal of Experimental Social Psychology, 38, 219–231.Colvin, C.R., Block, J., & Funder, D.C. (1995). Overly positive self-evaluationsand personality: Negative implications for mental health. Journal ofPersonality and Social Psychology, 68, 1152–1162.Corning, A.F. (2002). Self-esteem as a moderator between perceived discriminationand psychological distress among women. Journal of CounselingPsychology, 49, 117–126.Covington, M.J. (1989). Self-esteem and failure in schools: Analysis and policyimplications. In A.M. Mecca, N.J. Smelser, & J. Vasconcellos (Eds.),The social importance of self-esteem (pp. 72–124). Berkeley: Universityof California Press.Crockenberg, S., & Soby, B. (1989). Self-esteem and teenage pregnancy: Thesocial importance of self-esteem. Berkeley: University of California Press.Crocker, J., & Luhtanen, R. (1990). Collective self-esteem and ingroup bias.Journal of Personality and Social Psychology, 58, 60–67.Crocker, J., & Schwartz, I. (1985). Prejudice and ingroup favoritism in a minimalintergroup situation: Effects of self-esteem. Personality and SocialPsychology Bulletin, 11, 379–386.Crocker, J., Thompson, L.L., McGraw, K.M., & Ingerman, C. (1987). Downwardcomparison, prejudice, and evaluations of others: Effects of self-esteem andthreat. Journal of Personality and Social Psychology, 52, 907–916.Crocker, J., & Wolfe, C.T. (2001). Contingencies of self-worth. PsychologicalReview, 108, 593–623.Crooks, R., & Baur, K. (1999). Our sexuality (7th ed.). New York: Brooks/Cole.Crowne, D.P., & Marlowe, D. (1960). A new scale of social desirability independentof psychopathology. Journal of Consulting Psychology, 24, 349–354.Cvetkovich, G., & Grote, B. (1980). Psychosocial development and the socialproblem of teenage illegitimacy. In C. Chilman (Ed.), Adolescent pregnancyand childbearing (pp. 15–41). Washington, DC: U.S. GovernmentPrinting Office.Davies, J., & Brember, I. (1999). Reading and mathematics attainments andself-esteem in years 2 and 6—an eight-year cross-sectional study. EducationalStudies, 25, 145–157.Davis, I. (1988, January 22). Ministry for feeling good. London Times, p. 10.Dawes, R.M. (1994). House of cards: Psychology and psychotherapy built onmyth. New York: Free Press.DeLongis, A., Folkman, S., & Lazarus, R.S. (1988). The impact of daily stresson health and mood: Psychological and social resources as mediators.Journal of Personality and Social Psychology, 54, 486–495.DeNeve, K.M., & Cooper, H. (1998). The happy personality: A meta-analysis40 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.of 137 personality traits and subjective well-being. Psychological Bulletin,124, 197–229.DeWit, D.J., Offord, D.R., Sanford, M., Rye, B.J., Shain, M., & Wright, R.(2000). The effect of school culture on adolescent behavioural problems:Self-esteem, attachment to learning, and peer approval of deviance asmediating mechanisms. Canadian Journal of School Psychology, 16,15–38.Di Paula, A., & Campbell, J.D. (2002). Self-esteem and persistence in the faceof failure. Journal of Personality and Social Psychology, 83, 711–724.Diener, E., & Diener, M. (1995). Cross-cultural correlates of life satisfactionand self-esteem. Journal of Personality and Social Psychology, 68, 653–663.Diener, E., Wolsic, B., & Fujita, F. (1995). Physical attractiveness and subjectivewell-being. Journal of Personality and Social Psychology, 69, 120–129.Dion, K.K., & Dion, K.C. (1975). Self-esteem and romantic love. Journal ofPersonality, 43, 39–57.Dolcini, M.M., & Adler, N.E. (1994). Perceived competencies, peer group affiliation,and risk behavior among early adolescents. Health Psychology,13, 496–506.Ehrlich, H.J. (1973). The social psychology of prejudice. New York: Wiley.Emler, N. (2001). Self-esteem: The costs and consequences of low self-worth.York, England: York Publishing Services.Emmons, R.A. (1984). Factor analysis and construct validity of the NarcissisticPersonality Inventory. Journal of Personality Assessment, 48, 291–300.Epstein, R. (2001, January/February). The prince of reason. Psychology Today,34, 66–76.Fein, S., & Spencer, S.J. (1997). Prejudice as self-image maintenance: Affirmingthe self through derogating others. Journal of Personality and SocialPsychology, 73, 31–44.Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior.Reading, MA: Addison-Wesley.Fleming, J.S., & Courtney, B.E. (1984). The dimensionality of self-esteem: Hierarchicalfacet model for revised measurement scales. Journal of Personalityand Social Psychology, 46, 404–421.Forsyth, D.R., & Kerr, N.A. (1999, August). Are adaptive illusions adaptive?Poster presented at the annual meeting of the American PsychologicalAssociation, Boston, MA.Forthofer, M.S., Janz, N.K., Dodge, J.A., & Clark, N.M. (2001). Gender differencesin the associations of self-esteem, stress and social support withfunctional health status among older adults with heart disease. Journal ofWomen and Aging, 31, 19–37.French, S.A., Leffert, N., Story, M., Neumark-Sztainer, D., Hannan, P., & Benson,P.L. (2001). Adolescent binge/purge and weight loss behaviors: Associationswith developmental assets. Journal of Adolescent Health, 28,211–221.Frone, M.R. (2000). Interpersonal conflict at work and psychological outcomes:Testing a model among young workers. Journal of OccupationalHealth Psychology, 5, 246–255.Furnham, A., & Cheng, H. (2000). Lay theories of happiness. Journal of HappinessStudies, 1, 227–246.Gabriel, M.T., Critelli, J.W., & Ee, J.S. (1994). Narcissistic illusions in selfevaluationsof intelligence and attractiveness. Journal of Personality, 62,143–155.Gerrard, M., Gibbons, F.X., Reis-Bergan, M., & Russell, D.W. (2000). Selfesteem,self-serving cognitions, and health risk behavior. Journal of Personality,68, 1177–1201.Glendinning, A. (1998). Family life, health and lifestyles in rural areas: Therole of self-esteem. Health Education, 2, 59–68.Glendinning, A., & Inglis, D. (1999). Smoking behaviour in youth: The problemof low self-esteem? Journal of Adolescence, 22, 673–682.Gray-Little, B., Williams, V.S.L., & Hancock, T.D. (1997). An item responsetheory analysis of the Rosenberg Self-Esteem Scale. Personality and SocialPsychology Bulletin, 23, 443–451.Greenberg, J., Solomon, S., Pyszczynski, T., Rosenblatt, A., Burling, J., Lyon,D., Simon, L., & Pinel, E. (1992). Why do people need self-esteem? Convergingevidence that self-esteem serves an anxiety-buffering function.Journal of Personality and Social Psychology, 63, 913–922.Greenwald, A.G., & Farnham, S.D. (2000). Using the implicit association testto measure self-esteem and self-concept. Journal of Personality and SocialPsychology, 79, 1022–1038.Hansford, B.C., & Hattie, J.A. (1982). The relationship between self andachievement/performance measures. Review of Educational Research,52, 123–142.Harter, S. (1993). Causes and consequences of low self-esteem in children andadolescents. In R. Baumeister (Ed.), Self-esteem: The puzzle of low selfregard(pp. 87–116). New York: Plenum Press.Heatherton, T.F., & Polivy, J. (1992). Chronic dieting and eating disorders: Aspiral model. In J.H. Crowther, D.L. Tennenbaum, S.E. Hobfold, & M.A.Parris (Eds.), The etiology of bulimia nervosa: The individual and familialcontext (pp. 133–155). Washington, DC: Hemisphere.Heatherton, T.F., & Vohs, K.D. (2000). Interpersonal evaluations followingthreats to self: Role of self-esteem. Journal of Personality and SocialPsychology, 78, 725–736.Heine, S.H., Lehman, D.R., Markus, H.R., & Kitayama, S. (1999). Is there auniversal need for positive self-regard? Psychological Review, 106, 766–794.Hendrick, C., & Hendrick, S.S. (1986). A theory and method of love. Journalof Personality and Social Psychology, 50, 392–402.Hendrick, S.S., Hendrick, C., & Adler, N.L. (1988). Romantic relationships:Love, satisfaction, and staying together. Journal of Personality and SocialPsychology, 54, 980–988.Herer, E., & Holzapfel, S. (1993). The medical causes of infertility and theireffects on sexuality. Canadian Journal of Human Sexuality, 2, 113–120.Herold, E.S., Corbesi, B., & Collins, J. (1994). Psychosocial aspects of femaletopless behavior on Australian beaches. Journal of Sex Research, 31,133–142.Herold, E.S., & Goodwin, M.S. (1979). Self-esteem and sexual permissiveness.Journal of Clinical Psychology, 35, 908–912.Herold, E.S., & Way, L. (1983). Oral-genital sexual behavior in a sample ofuniversity females. Journal of Sex Research, 19, 327–338.Hill, S.Y., Shen, S., Lowers, L., & Locke, J. (2000). Factors predicting the onsetof adolescent drinking in families at high risk for developing alcoholism.Biological Psychiatry, 48, 265–275.Hollar, D.S., & Snizek, W.E. (1996). The influences of knowledge of HIV/AIDS and self-esteem on the sexual practices of college students. SocialBehavior and Personality, 24, 75–86.Howerton, D.L., Enger, J.M., & Cobbs, C.R. (1994). Self-esteem and achievementof at-risk adolescent Black males. Research in the Schools, 1, 23–27.Hurlbert, D.F., & Whittaker, K.E. (1991). The role of masturbation in maritaland sexual satisfaction: A comparative study of female masturbators andnonmasturbators. Journal of Sex Education and Therapy, 17, 272–282.Jackson, C., Henriksen, L., Dickinson, D., & Levine, D.W. (1997). The earlyuse of alcohol and tobacco: Its relation to children’s competence and parents’behavior. American Journal of Public Health, 87, 359–364.James, K. (1997). Worker social identity and health-related costs for organizations:A comparative study between ethnic groups. Occupational Health,2, 108–117.Jang, S.J., & Thornberry, T.P. (1998). Self-esteem, delinquent peers, and delinquency:A test of the self-enhancement hypothesis. American SociologicalReview, 63, 586–598.Janis, I.L. (1982). Groupthink: Psychological studies of policy decisions and fiascoes.Boston: Houghton Mifflin.Janis, I.L., & Field, P. (1959). Sex differences and personality factors related topersuasibility. In C. Hovland & I.L. Janis (Eds.), Personality and persuasibility(pp. 55–68, 300–302). New Haven, CT: Yale University Press.Janoff-Bulman, R. (1992). Shattered assumptions: Towards a new psychologyof trauma. New York: Free Press.Janoff-Bulman, R., & Brickman, P. (1982). Expectations and what people learnfrom failure. In N.T. Feather (Ed.), Expectations and actions: Expectancy-valuemodels in psychology (pp. 207–237). Hillsdale, NJ: Erlbaum.Jessor, S.L., & Jessor, R. (1975). Transition from virginity to nonvirginityamong youth: A social-psychological study over time. DevelopmentalPsychology, 11, 473–484.Joiner, T.E., Alfano, M.S., & Metalsky, G.I. (1992). When depression breedscontempt: Reassurance seeking, self-esteem, and rejection of depressedcollege students by their roommates. Journal of Abnormal Psychology,101, 165–173.Judge, T.A., & Bono, J.E. (2001). Relationship of core self-evaluations traits—self-esteem, generalized self-efficacy, locus of control, and emotional stability—withjob satisfaction and job performance: A meta-analysis. Journalof Applied Psychology, 86, 80–92.VOL. 4, NO. 1, MAY 2003 41

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-EsteemJudge, T.A., Erez, A., & Bono, J.E. (2002). Are measures of self-esteem, neuroticism,locus of control, and generalized self-efficacy indicators of acommon core construct? Journal of Personality and Social Psychology,83, 693–710.Kalil, A., & Kunz, J. (1999). First births among unmarried adolescent girls:Risk and protective factors. Social Work Research, 23, 197–208.Kaplan, H.B., Smith, P.B., & Pokorny, A.D. (1979). Psychosocial antecedentsof unwed motherhood among indigent adolescents. Journal of Youth andAdolescence, 8, 181–207.Keefe, K., & Berndt, T.J. (1996). Relations of friendship quality to self-esteemin early adolescence. Journal of Early Adolescence, 16, 110–129.Kendall-Tackett, K.A., Williams, L.M., & Finkelhor, D. (1993). Impact of sexualabuse on children: A review and synthesis of recent empirical studies.Psychological Bulletin, 113, 164–180.Kernis, M.H., Grannemann, B.D., & Barclay, L.C. (1989). Stability and levelof self-esteem as predictors of anger arousal and hostility. Journal of Personalityand Social Psychology, 56, 1013–1022.Kernis, M.H., & Waschull, S.B. (1995). The interactive roles of stability andlevel of self-esteem: Research and theory. In M.P. Zanna (Ed.), Advancesin experimental social psychology (Vol. 27, pp. 93–141). San Diego, CA:Academic Press.Kirkpatrick, L.A., & Ellis, B.J. (2001). An evolutionary-psychological perspectiveon self-esteem: Multiple domains and multiple functions. InG.J.O. Fletcher & M.S. Clark (Eds.), Blackwell handbook of social psychology:Vol. 2. Interpersonal processes (pp. 411–436). Oxford, England:Blackwell.Kirschbaum, C., Prussner, J.D., Stone, A.A., Federenko, I., Gaab, J., Lintz, D.,Schommer, H.J., & Hellhammer, D.H. (1995). Persistent high cortisol responsesto repeated psychology distress in a subpopulation of healthymen. Psychosomatic Medicine, 57, 469–474.Koval, J.J., & Pederson, L.L. (1999). Stress-coping and other psychosocial riskfactors: A model for smoking in grade 6 students. Addictive Behaviors,24, 207–218.Krueger, J. (1998). Enhancement bias in the description of self and others. Personalityand Social Psychology Bulletin, 24, 505–516.Kugle, C.L., Clements, R.O., & Powell, P.M. (1983). Level and stability ofself-esteem in relation to academic behavior of second graders. Journalof Personality and Social Psychology, 44, 201–207.Lakey, B., Tardiff, T.A., & Drew, J.B. (1994). Negative social interactions: Assessmentand relations to social support, cognition, and psychologicaldistress. Journal of Social and Clinical Psychology, 13, 42–62.Langer, L.M., & Tubman, J.G. (1997). Risky sexual behavior among substance-abusingadolescents: Psychosocial and contextual factors. AmericanJournal of Orthopsychiatry, 67, 315–322.Lazarus, R.S., & Folkman, S. (1984). Stress, appraisal, and coping. New York:Springer.Leary, M.R., & Baumeister, R.F. (2000). The nature and function of selfesteem:Sociometer theory. Advances in Experimental Social Psychology,32, 1–62.Leary, M.R., & Downs, D.L. (1995). Interpersonal functions of the self-esteemmotive: The self-esteem system as a sociometer. In M.R. Leary & D.L.Downs (Eds.), Efficacy, agency, and self-esteem (pp. 123–144). NewYork: Plenum Press.Leary, M.R., Tambor, E.S., Terdal, S.K., & Downs, D.L. (1995). Self-esteem asan interpersonal monitor: The sociometer hypothesis. Journal of Personalityand Social Psychology, 68, 518–530.LePine, J.A., & Van Dyne, L. (1998). Predicting voice behavior in workgroups. Journal of Applied Psychology, 83, 853–868.Lerner, R.M., & Galambos, N.L. (1998). Adolescent development: Challengesand opportunities for research, programs, and policies. Annual Review ofPsychology, 49, 413–446.Lewis, P.C., Harrell, J.S., Bradley, C., & Deng, S. (2001). Cigarette use in adolescents:The Cardiovascular Health in Children and Youth Study. Researchin Nursing and Health, 24, 27–37.Lobel, T.E., & Levanon, I. (1988). Self-esteem, need for approval, and cheatingbehavior in children. Journal of Educational Psychology, 80, 122–123.Lyubomirsky, S., & Lepper, H.S. (2002). What are the differences betweenhappiness and self-esteem? Unpublished manuscript, University of California,Riverside.Martin, J.I., & Knox, J. (1997). Self-esteem instability and its implications forHIV prevention among gay men. Health and Social Work, 22, 264–273.Maruyama, G., Rubin, R.A., & Kingsbury, G.G. (1981). Self-esteem and educationalachievement: Independent constructs with a common cause?Journal of Personality and Social Psychology, 40, 962–975.Matz, P.E., Foster, G.D., Faith, M.S., & Wadden, T.A. (2002). Correlates ofbody image dissatisfaction among overweight women seeking weightloss. Journal of Consulting and Clinical Psychology, 70, 1040–1044.McFarlin, D.B. (1985). Persistence in the face of failure: The impact of self-esteemand contingency information. Personality and Social PsychologyBulletin, 11, 153–163.McFarlin, D.B., Baumeister, R.F., & Blascovich, J. (1984). On knowing whento quit: Task failure, self-esteem, advice, and nonproductive persistence.Journal of Personality, 52, 138–155.McGee, R., & Williams, S. (2000). Does low self-esteem predict health compromisingbehaviours among adolescents? Journal of Adolescence, 23,569–582.Mecca, A.M., Smelser, N.J., & Vasconcellos, J. (Eds.). (1989). The social importanceof self-esteem. Berkeley: University of California Press.Metalsky, G.I., Joiner, T.E., Hardin, T.S., & Abramson, L.Y. (1993). Depressivereactions to failure in a naturalistic setting: A test of the hopelessnessand self-esteem theories of depression. Journal of Abnormal Psychology,102, 101–109.Miller, B.C., Christensen, C.R., & Olson, T.D. (1987). Adolescent self-esteem inrelation to sexual attitudes and behavior. Youth and Society, 18, 93–111.Miller, C.T., & Downey, K.T. (1999). A meta-analysis of heavyweight and selfesteem.Personality and Social Psychology Review, 3, 68–84.Mintz, L.B., & Betz, N.E. (1988). Prevalence and correlates of eating disorderedbehaviors among undergraduate women. Journal of CounselingPsychology, 35, 463–471.Moore, D., & Li, L. (1998). Prevalence and risk factors of illicit drug use bypeople with disabilities. American Journal on Addictions, 7, 93–102.Murphy, C.M., Stosny, S., & Morrel, T.M. (2001). Enhanced self-esteem is associatedwith violence reduction among domestic abusers. Unpublishedmanuscript, University of Maryland, Baltimore County, Baltimore.Murray, S.L., Rose, P., Bellavia, G., Holmes, J.G., & Kusche, A. (2002). Whenrejection stings: How self-esteem constrains relationship-enhancementprocesses. Journal of Personality and Social Psychology, 83, 556–573.Murrell, S.A., Meeks, S., & Walker, J. (1991). Protective functions of healthand self-esteem against depression in older adults facing illness or bereavement.Psychology and Aging, 6, 352–360.National Association for Self-Esteem. (2000). Masters coalition—a work inprogress. Retrieved January 25, 2002, from http://www.self-esteem-nase.org/masters.shtmlNeumark-Sztainer, D., Story, M., French, S.A., & Resnick, M.D. (1997). Psychosocialcorrelates of health compromising behaviors among adolescents.Health Education Research, 12, 37–52.Nirkko, O., Lauroma, H.J., Siltanen, P., Tuominen, H., & Vanhala, K. (1982).Psychological risk factors related to coronary heart disease: Prospectivestudies among policemen in Helsinki. Acta Medica Scandinavica—Supplementum,660, 137–146.Oates, G.L. (1997). Self-esteem enhancement through fertility? Socioeconomicprospects, gender, and mutual influence. American SociologicalReview, 62, 965–973.Olweus, D. (1990). Bullying among school children. In K. Hurrelmann & F.Loesel (Eds.), Health hazards in adolescence (pp. 259–297). Berlin, Germany:Walter De Gruyter.Olweus, D. (1994). Bullying at school: Long-term outcomes for the victimsand an effective school-based intervention program. In R. Huesmann(Ed.), Aggressive behavior: Current perspectives (pp. 97–130). NewYork: Plenum Press.Ornish, D. (1998). Love and survival. New York: HarperCollins.Ortiz, V., & Volloff, W. (1987). Identification of gifted and accelerated Hispanicstudents. Journal for the Education of the Gifted, 11, 45–55.Paul, C., Fitzjohn, J., Herbison, P., & Dickson, N. (2000). The determinants ofsexual intercourse before age 16. Journal of Adolescent Health, 27, 136–147.Paulhus, D.L. (1998). Interpersonal and intrapsychic adaptiveness of trait selfenhancement:A mixed blessing? Journal of Personality and Social Psychology,74, 1197–1208.Paulhus, D.L. (2002). Socially desirable responding: The evolution of a construct.In H.I. Braun & D.N. Jackson (Eds.), The role of constructs in psychologicaland educational measurement (pp. 37–48). Mahwah, NJ: Erlbaum.Pederson, L.L., Koval, J.J., McGrady, G.A., & Tyas, S.L. (1998). The degree42 VOL. 4, NO. 1, MAY 2003

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTR.F. Baumeister et al.and type of relationship between psychosocial variables and smoking statusfor students in grade 8: Is there a dose-response relationship? PreventativeMedicine, 27, 337–347.Perez, R.C. (1973). The effect of experimentally induced failure, self-esteemand sex on cognitive differentiation. Journal of Abnormal Psychology,81, 74–79.Plotnick, R.D. (1992). The effects of attitudes on teenage premarital pregnancyand its resolution. American Sociological Review, 57, 800–811.Poikolainen, K., Tuulio-Henriksson, A., Aalto-Setälä, T., Marttunen, M., &Lönnqvist, J. (2001). Predictors of alcohol intake and heavy drinking inearly adulthood: A 5-year follow-up of 15-19 year old Finnish adolescents.Alcohol and Alcoholism, 36, 85–88.Polivy, J., Heatherton, T.F., & Herman, C.P. (1988). Self-esteem, restraint, andeating behavior. Journal of Abnormal Psychology, 97, 354–356.Pottebaum, S.M., Keith, T.Z., & Ehly, S.W. (1986). Is there a causal relationbetween self-concept and academic achievement? Journal of EducationalResearch, 79, 140–144.Pruessner, J.C., Hellhammer, D.H., & Kirschbaum, D. (1998). Low self-esteem,induced failure and the adrenocortical stress response. Personality andIndividual Differences, 27, 477–489.Ralph, J.A., & Mineka, S. (1998). Attributional style and self-esteem: The predictionof emotional distress following a midterm exam. Journal of AbnormalPsychology, 107, 203–215.Raskin, R.N., & Hall, C.S. (1981). The Narcissistic Personality Inventory: Alternativeform reliability and further evidence of construct validity. Journalof Personality Assessment, 45, 159–162.Rhodewalt, F., Madrian, J.C., & Cheney, S. (1998). Narcissism, self-knowledgeorganization, and emotional reactivity: The effect of daily experienceson self-esteem and affect. Personality and Social PsychologyBulletin, 24, 75–87.Rhodewalt, F., & Morf, C.C. (1995). Self and interpersonal correlates of theNarcissistic Personality Inventory: A review and new findings. Journal ofResearch in Personality, 29, 1–23.Robbins, C., Kaplan, H.B., & Martin, S.S. (1985). Antecedents of pregnancyamong unmarried adolescents. Journal of Marriage and the Family, 47,567–583.Roberts, J.E., & Monroe, S.M. (1992). Vulnerable self-esteem and depressivesymptoms: Prospective findings comparing three alternative conceptualizations.Journal of Personality and Social Psychology, 62, 804–812.Roberts, J.E., & Monroe, S.M. (1994). A multidimensional model of self-esteemin depression. Clinical Psychology Review, 14, 161–181.Roberts, J.E., & Monroe, S.M. (1999). Vulnerable self-esteem and social processesin depression: Toward an interpersonal model of self-esteem regulation.In T.E. Joiner, Jr., & J.C. Coyne (Eds.), The interactional nature ofdepression (pp. 149–188). Washington, DC: American Psychological Association.Robins, R.W., Hendin, H.M., & Trzesniewski, K.H. (2001). Measuring globalself-esteem: Construct validation of a single-item measure and theRosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin,27, 151–161.Robinson, N.S., Garber, J., & Hilsman, R. (1995). Cognitions and stress: Directand moderating effects on depressive versus externalizing symptomsduring the junior high school transition. Journal of Abnormal Psychology,104, 453–463.Roese, N.J., & Pennington, G.L. (2002). Ego-protective judgments of event impact:The person/group discrimination discrepancy and beyond. Unpublishedmanuscript, Northwestern University, Evanston, IL.Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ:Princeton University Press.Rosenberg, M., Schooler, C., & Schoenbach, C. (1989). Self-esteem and adolescentproblems: Modeling reciprocal effects. American SociologicalReview, 54, 1004–1018.Rosenberg, M., Schooler, C., Schoenbach, C., & Rosenberg, F. (1995). Globalself-esteem and specific self-esteem: Different concepts, different outcomes.American Sociological Review, 60, 141–156.Rosenthal, R., & Jacobson, L. (1968). Pygmalion in the classroom: Teacher expectationand pupils’ intellectual development. New York: Holt.Rubin, R.A., Dorle, J., & Sandidge, S. (1977). Self-esteem and school performance.Psychology in the Schools, 14, 503–507.Rusbult, C.E., Morrow, G.D., & Johnson, D.J. (1987). Self-esteem and problem-solvingbehaviour in close relationships. British Journal of SocialPsychology, 26, 293–303.Safer, D.L., Lively, T.J., Telch, C.F., & Agras, W.S. (2002). Predictors of relapsefollowing successful dialetical behavior therapy for binge eatingdisorder. International Journal of Eating Disorders, 32, 155–162.Salmivalli, C., Kaukiainen, A., Kaistaniemi, L., & Lagerspetz, K.M.J. (1999).Self-evaluated self-esteem, peer-evaluated self-esteem, and defensiveegotism as predictors of adolescents’ participation in bullying situations.Personality and Social Psychology Bulletin, 25, 1268–1278.Sandelands, L.E., Brockner, J., & Glynn, M.A. (1988). If at first you don’t succeed,try, try again: Effects of persistence-performance contingencies,ego involvement, and self-esteem on task persistence. Journal of AppliedPsychology, 73, 208–216.Scheier, L.M., Botvin, G.J., Griffin, K.W., & Diaz, T. (2000). Dynamic growthmodels of self-esteem and adolescent alcohol use. Journal of Early Adolescence,20, 178–209.Scheirer, M.A., & Kraut, R.E. (1979). Increased educational achievement viaself-concept change. Review of Educational Research, 49, 131–150.Schneider, D.J., & Turkat, D. (1975). Self-presentation following success orfailure: Defensive self-esteem models. Journal of Personality, 43, 127–135.Sedikides, C. (1993). Assessment, enhancement, and verification determinantsof the self-evaluation process. Journal of Personality and Social Psychology,65, 317–338.Seeman, T.E., Berkman, L.F., Gulanski, B.I., Robbins, R.J., Greenspan, S.L.,Charpentier, P.A., & Rowe, J.W. (1995). Self-esteem and neuroendocrineresponse to challenge: MacArthur studies of successful aging. Journal ofPsychosomatic Research, 39, 69–84.Shackelford, T.K. (2001). Self-esteem in marriage. Personality and IndividualDifferences, 30, 371–390.Sharp, M.J., & Getz, J.G. (1996). Substance use as impression management.Personality and Social Psychology Bulletin, 22, 60–67.Shrauger, J.S., & Sorman, P.B. (1977). Self-evaluations, initial success andfailure, and improvement as determinants of persistence. Journal of Consultingand Clinical Psychology, 45, 784–795.Simon, W.E., & Simon, M.G. (1975). Self-esteem, intelligence, and standardizedacademic achievement. Psychology in the Schools, 12, 97–100.Sinha, R.R., & Krueger, J. (1998). Idiographic self-evaluation and bias. Journalof Research in Personality, 32, 131–155.Skaalvik, E.M., & Hagtvet, K.A. (1990). Academic achievement and self-concept:An analysis of causal predominance in a developmental perspective.Journal of Personality and Social Psychology, 58, 292–307.Slee, P.T., & Rigby, K. (1993). Australian school children’s self appraisal of interpersonalrelations: The bullying experience. Child Psychiatry and HumanDevelopment, 23, 273–282.Smelser, N.J. (1989). Self-esteem and social problems: An introduction. InA.M. Mecca, N.J. Smelser, & J. Vasconcellos (Eds.), The social importanceof self-esteem (pp. 1–23). Berkeley: University of CaliforniaPress.Smith, G.E., Gerrard, M., & Gibbons, F.X. (1997). Self-esteem and the relationbetween risk behavior and perceptions of vulnerability to unplannedpregnancy in college women. Health Psychology, 16, 137–146.Stokes, J.P., Damon, W., & McKirnan, D.J. (1997). Predictors of movement towardhomosexuality: A longitudinal study of bisexual men. Journal ofSex Research, 34, 304–312.Strouse, J.S., & Buerkel-Rothfuss, N.L. (1987). Media exposure and the sexualattitudes and behaviors of college students. Journal of Sex Education andTherapy, 13, 43–51.Stucke, T.S., & Sporer, S.L. (2002). When a grandiose self-image is threatened:Narcissism and self-concept clarity as predictors of negative emotionsand aggression following ego-threat. Journal of Personality, 70,509–532.Swann, W.B., Jr., Stein-Seroussi, A., & Giesler, R.B. (1992). Why people selfverify.Journal of Personality and Social Psychology, 62, 392–401.Taylor, S.E. (1983). Adjustment to threatening events: A theory of cognitiveadaptation. American Psychologist, 38, 1161–1173.Tharenou, P. (1979). Employee self-esteem: A review of the literature. Journalof Vocational Behavior, 15, 316–346.Tice, D.M., Butler, J.L., Muraven, M.B., & Stillwell, A.M. (1995). When modestyprevails: Differential favorability of self-presentation to friends andstrangers. Journal of Personality and Social Psychology, 69, 1120–1138.Trope, Y. (1986). Self-enhancement and self-assessment in achievement behavior.In Y. Trope (Ed.), Handbook of motivation and cognition: Foundationsof social behavior (pp. 350–378). New York: Guilford Press.VOL. 4, NO. 1, MAY 2003 43

PSYCHOLOGICAL SCIENCE IN THE PUBLIC INTERESTBenefits of Self-EsteemTrzesniewski, K.H., Donnellan, M.B., Robins, R.W., Moffitt, T.E., & Caspi, A.(2002, February). Do juvenile delinquents have high or low self-esteem?Paper presented at the annual meeting of the Society for Personality andSocial Psychology, Savannah, GA.Twenge, J.M., & Campbell, W.K. (2001). Age and birth cohort differences inself-esteem: A cross-temporal meta-analysis. Personality and Social PsychologyReview, 5, 321–344.van-der-Ham, T., van-Strein, D.C., & van-Engeland, H. (1998). Personalitycharacteristics predict outcomes of eating disorders in adolescents: A 4-yearprospective study. European Child and Adolescent Psychiatry, 7, 79–84.Vernon, M.E.L., Green, J.A., & Frothingham, T.E. (1983). Teenage pregnancy:A prospective study of self-esteem and other sociodemographic factors.Pediatrics, 72, 632–635.Vohs, K.D., Bardone, A.M., Joiner, T.E., Jr., Abramson, L.Y., & Heatherton,T.F. (1999). Perfectionism, perceived weight status, and self-esteem interactto predict bulimic symptoms: A model of bulimic symptom development.Journal of Abnormal Psychology, 108, 695–700.Vohs, K.D., & Heatherton, T.F. (2001). Self-esteem and threats to self: Implicationsfor self-construals and interpersonal perceptions. Journal of Personalityand Social Psychology, 81, 1103–1118.Vohs, K.D., & Heatherton, T.F. (in press). The effects of self-esteem and egothreat on interpersonal appraisals of men and women: A naturalisticstudy. Personality and Social Psychology Bulletin.Vohs, K.D., Heatherton, T.F., Baumeister, R.F., Bardone, A.M., Abramson,L.Y., & Joiner, T.E., Jr. (2002). A new psychological model of the developmentof bulimic symptoms. Unpublished manuscript, University ofUtah, Salt Lake City.Vohs, K.D., Voelz, Z.R., Pettit, J.W., Bardone, A.M., Katz, J., Abramson, L.Y.,Heatherton, T.F., & Joiner, T.E., Jr. (2001). Perfectionism, body dissatisfaction,and self-esteem: An interactive model of bulimic symptom development.Journal of Social and Clinical Psychology, 20, 476–496.Waldo, C.R., Hesson-McInnis, M.S., & D’Augelli, A.R. (1998). Antecedentsand consequences of victimization of lesbian, gay, and bisexual youngpeople: A structural model comparing rural university and urban samples.American Journal of Community Psychology, 26, 307–334.Wallace, H.M., & Baumeister, R.F. (2002). The performance of narcissistsrises and falls with perceived opportunity for glory. Journal of Personalityand Social Psychology, 82, 819–834.Walsh, A. (1991). Self-esteem and sexual behavior: Exploring gender differences.Sex Roles, 25, 441–450.West, P., & Sweeting, H. (1997). “Lost souls” and “rebels”: A challenge to theassumption that low self-esteem and unhealthy lifestyles are related.Health Education, 5, 161–167.Whisman, M.A., & Kwon, P. (1993). Life stress and dysphoria: The role ofself-esteem and hopelessness. Journal of Personality and Social Psychology,65, 1054–1060.Williams, G.J., Power, K.G., Millar, H.R., Freeman, C.P., Yellowlees, A.,Dowds, T., Walker, M., Campsie, L., MacPherson, F., & Jackson, M.A.(1993). Comparison of eating disorders and other dietary/weightgroups on measures of perceived control, assertiveness, self-esteem, andself-directed hostility. International Journal of Eating Disorders, 14,27–32.Wills, T.A. (1981). Downward comparison principles in social psychology.Psychological Bulletin, 90, 245–271.Wills, T.A. (1994). Self-esteem and perceived control in adolescent substanceuse: Comparative tests in concurrent and prospective analyses. Psychologyof Addictive Behaviors, 8, 223–234.Winegar, K. (1990, November 27). Self-esteem is healthy for society. Star Tribune[Minneapolis, MN], pp. 1E–2E.Wylie, R.C. (1979). The self-concept: Vol. 2. Theory and research on selectedtopics. Lincoln: University of Nebraska Press.Zimmerman, M.A., Copeland, L.A., Shope, J.T., & Dielman, T.E. (1997). Alongitudinal study of self-esteem: Implications for adolescent development.Journal of Youth and Adolescence, 26, 117–141.44 VOL. 4, NO. 1, MAY 2003