BOLSTERING SELF–ESTEEMFORSYTH ET AL.
ATTEMPTING TO IMPROVE THE ACADEMIC
PERFORMANCE OF STRUGGLING
COLLEGE STUDENTS BY BOLSTERING
AN INTERVENTION THAT BACKFIRED
DONELSON R. FORSYTH
University of Richmond
NATALIE K. LAWRENCE
James Madison University
JENI L. BURNETTE
University of Richmond
ROY F. BAUMEISTER
Florida State University
Theory and prior research suggest that (a) a positive sense of self–worth and (b) per-
ceived control over one’s outcomes facilitate constructive responses to negative
outcomes. We therefore predicted that encouraging students to maintain their
sense of self–worth and/or construe their academic outcomes as controllable
would promote achievement. In a field experiment, low–performing students in a
psychology class were randomly assigned to receive, each week, review questions,
review questions plus self–esteem bolstering, or review questions plus exhortations
to assume responsibility and control. Contrary to predictions, the D and F students
got worse as a result of self–esteem bolstering and students in the other conditions
did not change. These findings raise ethical and practical questions about the wide-
spread practice of bolstering self–esteem in the hope of improving academic
Journal of Social and Clinical Psychology, Vol. 26, No. 4, 2007, pp. 447–459
Portions of his manuscript were reported at the Annual Meetings of the American Psy-
Correspondence concerning this article should be addressed to Donelson R. Forsyth,
Jepson School of Leadership Studies, University of Richmond, Richmond, VA 23173.
People may claim they are inspired by the Delphic command to “know
thyself,” but the self they tend to seek is one that is uniquely superior to
others and in control of important outcomes (Dunning, Heath, & Suls,
2004). Of the self’s many motives, maintaining adequately high self–es-
teem and achieving a sense of control over the environment are among
the most commonly observed in Western civilization (e.g., Baumeister,
1998; Leary, 2004). Both these needs have captured the imagination of re-
searchers as important keys to psychological functioning. Students’ suc-
cess in school, in particular, is widely thought to suffer if either
self–esteem or a sense of control is lost.
The present research was a field experiment that sought to improve
the performance of below–average students in a psychology course by
bolstering either their self–esteem or their sense of control. It was guided
by the sincerely optimistic belief that these interventions would succeed
and could then form the model for both further research and practical in-
terventions. There would be immense social value if one could establish
social interventions that improved learning and test–taking. Folk wis-
dom and some findings indicate that school performance is mainly a
function of intelligence and studying (Sternberg, 1998), but insofar as in-
telligence is hard to change, and studying can be tedious and aversive,
other ways of improving educational outcomes would be welcome.
Bolstering self–esteem has long seemed a promising and appealing
way to improve student performance. Researchers began to find posi-
tive correlations between self–esteem and school performance (grades
and test scores) in the 1960s, and by 1979 Wylie concluded that the link
was sufficiently well–replicated to be considered a proven fact, although
the correlations were typically relatively modest (Valentine, DuBois, &
Cooper, 2004). Socially concerned persons speculated that the disturb-
ingly low academic performance of some stigmatized minority groups
might be caused by their low self–esteem resulting from society’s mes-
sages of inferiority. Sound theoretical arguments further supported the
idea that self–esteem contributes to improved performance, for high
self–esteem should lead to setting higher goals and the confidence
needed to bounce back from initial failure with renewed effort (see
Bachman & O’Malley, 1977; Di Paula & Campbell, 2002; McFarlin &
Baumeister, 1984). The view that confidence promotes effort and
thereby brings success is deeply rooted in American culture.
Hence, it is hardly surprising that many schools began using self–es-
teem bolstering programs in the justifiable hope that this would lead to
better school performance. The self–esteem programs proved popular
with both teachers and students, possibly because exercises in
self–praise and reciprocal flattery are somewhat more enjoyable uses of
class time than, say, doing arithmetic. A 2006 Google.com search for “el-
448 FORSYTH ET AL.
ementary school mission statement self–esteem” found 308,000 web
sites (Twenge, 2006). Generally those mission statements say that elevat-
ing self–esteem of students is a primary goal to be pursued in advance of
academic performance rather than as a result.
In retrospect, there were some reasons to doubt that bolstering self–es-
teem would bear fruit in terms of improved test scores. The well–estab-
lished positive correlation between self–esteem and academic perfor-
mance that Wylie (1979) affirmed was not necessarily due to the causal
effects of self–esteem. Rather, good grades in school were more likely
the cause of increases in self–esteem (alongside third–variable causes;
see Bachman & O’Malley, 1977, 1986; Maruyama, Rubin, & Kingsbury,
1981). The minority groups turned out not to suffer uniformly from low
self–esteem as had been supposed (Crocker & Major, 1989; Twenge &
Crocker, 2002), and Sommers (1995) pointed out that the group data on
school performance of White and Black boys and girls suggested an in-
verse correlation with self–esteem, such that White girls had the lowest
average self–esteem and the best school records, while Black boys
scored highest on self–esteem but lowest in school performance. As for
interventions aimed specifically at boosting self–esteem, an early litera-
ture review by Scheirer and Kraut (1979) reached the pessimistic conclu-
sion that these interventions do not succeed at improving school perfor-
mance. Similarly, Valentine and his colleagues (2004), in their
meta–analytic synthesis of studies that measured self–esteem and per-
formance at several points in time, concluded the evidence did not
support “interventions that are aimed solely at improving students’
views of themselves” (p. 129).
The lack of evidence about the academic benefits of bolstering self–es-
teem, despite the widespread school programs, may also reflect meth-
odological drawbacks. The very belief in the value of self–esteem has
made it seem cruel and unethical to withhold these benefits from some
students, so most interventions lack any sort of control group that could
be used as a comparison baseline to assess the value of boosting self–es-
teem. The present study was designed with random assignment of indi-
vidual participants to self–esteem boosting or not, and we suspect that
almost no elementary or secondary school self–esteem program is ad-
ministered on that basis. This study offers an opportunity to assess the
effects of self–esteem bolstering on academic performance with an
Personal control was a second focus of the present investigation. As
with self–esteem, we had ample theoretical and empirical justification
for predicting that bolstering a sense of control would improve test per-
formance. In school, control means taking responsibility for one’s work
and attributing performance to inner causes, especially effort. Prior evi-
BOLSTERING SELF–ESTEEM 449
dence indicates that a sense of being unable to control one’s outcomes
can be debilitating, even paralyzing, in the form of learned helplessness
(Seligman, 1975). Students who regularly attribute their grades to fac-
tors they control are more successful than those who think they do not
control their academic outcomes (Diener & Dweck, 1980). When Dweck
(1975) trained children to attribute their failures to lack of effort—an in-
ternal, controllable cause—they showed improved persistence after ini-
tial failure. Wilson and Linville (1982, 1985) used attributional training
to convince college students that most of their peers managed to im-
prove their grades over the course of their college careers. Relative to
controls, these students were less likely to drop out and their grades im-
proved. Noel, Forsyth, and Kelley (1987) similarly found that students
who were failing a psychology class responded well when their attribu-
tions were shifted away from external, uncontrollable factors (such as
“difficult test”) to internal, controllable causes (such as effort and moti-
vation). This attributional bolstering of the sense of personal control and
responsibility led to higher grades at the end of the course, as compared
to a control group.
The present study targeted all students who received C, D, or F on the
first examination in a large psychology course. By random assignment,
some received weekly messages for the rest of the term aimed at bolster-
ing their self–esteem. A second group received messages aimed at bol-
stering the sense of internal control and personal responsibility. A third
group received no such intervention, but all three groups did receive
weekly review questions as study aids. This design is consistent with the
procedures used in many school self–esteem programs, including
for–profit interventions. These programs usually include study skills
training, seemingly in implicit acknowledgment that bolstering self–es-
teem alone is not enough. The present investigation offered a rare
chance to separate the benefits of self–esteem bolstering from study aids,
even while using both together in the crucial condition.
We undertook this study expecting that academic performance would
improve from both the self–esteem and the internal/control interven-
tions, but that either one might prove more effective than the other. We
felt it was ethically acceptable to include a control group that did not re-
ceive any such intervention because they were essentially receiving the
same treatment that most students in most classes received, and so no
group was put at a disadvantage relative to every day life. Moreover, if
we could establish that one or possibly both interventions reliably im-
proved learning and/or test performance, then these findings could
benefit many students. Our a priori hypothesizing was mainly divided
as to whether the self–esteem or the internal/control intervention would
prove more successful at improving grades.
450 FORSYTH ET AL.
All students who received a C, D, or F (thus below 80% correct) on the
first major examination in a large psychology class, and who had stu-
dent email accounts listed on the class listserve were invited to partici-
pate in return for partial course credit. Thus, 141 of 305 students received
invitations, 51 failed to respond. The remaining 90 (68 women) took
part. They were tracked in two groups, based on having earned a C (N =
45) versus a D or F (N = 45). Four students withdrew before the semester
ended, leaving a final sample of 86.
The entire study was conducted via electronic mail. Participants were
initially contacted via the student email accounts and invited to take part
in a study concerned with “communication and the use of email.” Partic-
ipants were told that the purpose of the emails was to provide them with
information about academic performance. Each week they would re-
ceive a message containing a review question pertaining to the material
covered in class, and they were required to acknowledge receipt of the
email within 24 hours to qualify for their credit as participants.
We then randomly assigned students into one of 3 conditions:
no–message control, internal/control, and self–esteem bolstering. Par-
ticipants in the control condition received only the review question each
week. Participants in the internal/control condition also received the re-
view question, but it was accompanied by messages (based on those
used by Noel et al., 1987) that encouraged them to take responsibility for
their performance in the course. The first message, for example, stated:
Past research suggests that when students get back their tests, they tend
to blame poor scores on external factors: they say things like “the test
was too hard,” or “the prof didn’t explain that,” or “the questions are
too picky.” Other studies suggest, though, that students who take re-
sponsibility for their grades not only get better grades, but they also
learn that they, personally, can control the grades they get . . . Bottom
line: Take personal control of your performance.
Students in the self–esteem bolstering condition also received the re-
view questions, but their email messages stressed the importance of
maintaining high self–esteem. The first message, for example, stated:
Past research suggests that when students get back their tests, they tend
BOLSTERING SELF–ESTEEM 451
to lose confidence: they say things like “I can’t do this,” or “I’m worth-
less,” or “I’m not as good as other people in college.” Other studies sug-
gest, though, that students who have high self–esteem not only get
better grades, but they remain self–confident and assured . . . Bottom
line: Hold your head—and your self–esteem—high.
We sent students an email message each week for a total of 6 weeks.
They received the last message the week before they completed the final
examination in the class, which we used as a measure of the impact of
each type of message on their academic performance.
As a manipulation check, the last email message also asked partici-
pants to indicate their degree of agreement, on a 5–point scale that
ranged from Agree Strongly (5) to Disagree Strongly (1), with 4 statements
pertaining to control and self–esteem. We averaged the items “I am in
control of my grades in Psychology 101” and “There are things that I can
do to control my grade in Psychology 101” to obtain an index of control
perceptions and the items “I feel good about myself as a student in Psy-
chology 101” and “I feel good about myself in general” to form a index of
self–esteem. Seventeen students did not provide responses to these
The average grade on the final examination, across the entire class of 255,
was 69.0%, a decrease from the first test mean of 74.8%. The students
who participated in the study were all ones who had earned Cs or lower
on the first test, and their grades declined from 67.7% to 63.5%. Exam
scores did not differ by sex; F(1, 84) = 0.44, ns.
The main dependent measure was score on the final examination, mea-
sured as percentage correct. We examined these scores using a 2 (Test 1
Grade: C vs. D/F) ×3 (Condition: No–Message Control, Internal/Con-
trol, and Self–esteem Bolstering) ×2 (Sex) ×2 (Test: Test 1 vs. Final)
mixed ANOVA, with repeated measures on the final factor. Two effects
that emerged in this analysis, the main effect of time, F(1, 74) = 12.40, p<
.01, η2= .14 and the interaction of time and condition, F(2, 74) = 3.08, p=
.05, η2= .08, were qualified by the 3–way interaction of time, condition,
and grade graphed in Figure 1; F(2, 74) = 3.10, p= .05, η2= .08. The pat-
tern of students’ grades was the opposite of our predictions, and in fact
D and F students in the self–esteem bolstering condition showed a sub-
stantial drop in grades from the midterm (57% correct) to 38% on the fi-
452 FORSYTH ET AL.
nal. This drop in scores was significant both statistically (p<.05) and
academically (dropping from the borderline for passing into the low F
range). The grades for C students declined somewhat from the first test
to the final, but the decrements did not vary systematically as a function
of message condition. The D and F students in the no–message control
condition and in the internal/controllability condition did not change.
BOLSTERING SELF–ESTEEM 453
FIGURE 1. Mean examination scores for C and D/F students in the Self-es-
teem (SE) bolstering condition, the Internal/Controllable Condition, and the
SELF–ESTEEM AND CONTROL
A 2 (Test 1 Grade: C grade vs. D/F grade) ×3 (Condition: Control, Inter-
nal/Control, and Self–esteem) ×2 (Sex) ANOVA of the control and
self–esteem indexes yielded limited support for the effectiveness of the
manipulations. Given the relatively high rate of non–responses on the fi-
nal questionnaire and the possible unreliability of these 2–item scales,
we suggest these effects be interpreted cautiously.
Self–esteem. Treatment condition had only a marginal effect on the two
items measuring self–esteem; F(2, 57) = 2.32 p<.11, η2= .08. However,
Duncan’s New Multiple Range Test indicated that students in the
self–esteem condition (M = 4.3) reported more positive self–evaluations
than students in the internal/control condition (M = 3.7), while the con-
trol condition (M = 3.9) fell in between and did not differ from either.
Moreover, these findings may reflect a restriction in range caused by a
ceiling effect: 70% of respondents in the self–esteem bolstering condition
and 50% of respondents in the other two conditions gave themselves the
highest possible rating on these items.1
Control. The only significant effect in the ANOVA of the control index
was the main effect of condition; F(2, 57) = 3.58 p<.05, η2= .11. Duncan’s
New Multiple Range Test indicated that students in the self–esteem con-
dition (M = 4.8) reported higher feelings of control than students in the
no–message condition (M = 4.3). The mean for the internal/control con-
dition (M = 4.5) fell intermediate and did no differ from either.
Correlational analyses. Sense of control and self–esteem were positively
correlated; r(69) = .24. Only self–esteem, however, was positively corre-
lated with both first exam grade, r(69) = .34 and final exam grade, r(69) =
.26, ps<.05. When the latter correlation was corrected for the former, it be-
came nonsignificant; r(66)= .09, ns. These results replicate earlier findings
that high self–esteem is correlated with better grades but not because of
any causal influence of self–esteem. That is, having high self–esteem after
our interventions predicted getting better grades on the exam, but this
was due to having gotten better grades on the earlier exam.
The present study presented an unusual opportunity to test the aca-
demic impact of bolstering self–esteem or a sense of control in a field ex-
454 FORSYTH ET AL.
1. The interaction of sex with grade on Test 1, F(1, 57) = 4.17 p<.05, 02= .07, on this item
indicated that men and women who had Cs on the first test reported nearly identical
self–esteem; 4.08 and 4.07. But among students who received Ds and Fs on the first test,
men reported much higher levels of self–esteem than women; 4.3 and 3.5, respectively.
periment design that included random assignment to conditions and
actual grades. We unexpectedly discovered that bolstering self–esteem
led to poorer performance, especially among the weaker (D and F) stu-
dents. Students with low grades on the first examination who then re-
ceived messages aimed at bolstering their self–esteem performed worse
afterwards. The decline in performance was meaningful in absolute
terms, going from borderline passing to a low F, and also in relative
terms, being significantly worse than the no–message control group and
the responsibility attribution group, both of which did not change. The C
students who received the self–esteem bolstering messages likewise got
worse, although C students in the other conditions also showed some
decline. The attempt to increase students’ feelings of responsibility and
control over their outcomes also failed to help them improve their
grades. Taylor and Brown (1988) suggested that a sense of control helps
“make each individual’s world a warmer and more active and beneficial
place in which to live” (p. 205), but our participants, who were reminded
to avoid externalizing the blame for their failures, did not improve their
scores over time.
Why did these interventions fail? Looking first at self–esteem, we col-
lected these data before the publication of Baumeister, Campbell,
Krueger, and Vohs’s (2003) and Valentine, DeBois, and Cooper’s (2004)
reviews of the self–esteem literature. Even though those reviews ques-
tion the beneficial effects of raising self–esteem, the negative effects of
bolstering self–esteem we report here are rare ones. Baumeister and his
colleagues cautioned that boosting self–evaluation independent of ac-
tual performance might encourage people to adopt a cavalier, defensive
attitude toward external demands and criteria. In the present context, it
is conceivable that the messages aimed at bolstering self–esteem some-
how convinced participants to think well of themselves regardless of
their academic efforts and outcomes. Indeed, weak students may main-
tain self–esteem best by withdrawing effort and minimizing the degree
to which their self–esteem is contingent on good grades (see Crocker &
Park, 2004; Forsyth, 1986). In the present context, such withdrawal
would most likely have had an adverse impact on studying and other
forms of effort, possibly leading to the poor performance on the final
Turning to a sense of control, recent reviews have raised questions
about the value of unrealistically positive views of the self, including ex-
aggerated feelings of control (Dunning et al., 2004). Even though a sense
of control may provide a buffer against the motivational side–effects of
helplessness, feelings of unrealistic optimism can lead to complacency
rather than active coping. When the students were assured that they,
themselves, controlled their outcomes, they may not engage in the types
BOLSTERING SELF–ESTEEM 455
of behaviors that would have improved their grades in the class, such as
studying, attending lectures, and completing assignments. Even though
prior studies using similar methods have successfully manipulated
sense of control, the use of emailed messages may have been less effec-
tive in effecting a change in sense of control. Recent findings, too, sug-
gest that a sense of controllability, independent of context or domain,
may not facilitate achievement in a specific setting. Self–efficacy theory,
for example, recommends using interventions that bolster confidence
with regard to the specific intellectual skills and learning activities
needed to perform well at the course’s tasks instead of inducing a
generalized sense of control (Lodewyk & Winne, 2005).
The present findings raise obstacles for further research. Self–esteem
bolstering interventions are widespread in the United States, as we
noted in the introduction, and for that reason deserve careful scrutiny. If
they are causing students with low grades to do worse in school—as the
present findings indicated, and that grade inflation could conceal in
most school settings—then it seems imperative to conduct careful, rigor-
ous tests. Longitudinal survey studies like those reviewed by Valentine
et al. (2004) provide data on large samples of students in school settings,
but random assignment to experimental conditions enables causal hy-
potheses to be tested reliably with substantially smaller samples. Given
the pragmatic difficulties of withholding treatments from students who
are thought likely to benefit from them, few such studies are done, and
so the published literature is largely devoid of studies like the present
one. Moreover, it may be unethical to conduct properly designed studies
in which one treatment has a chance of causing students to get worse
grades than they would otherwise. The present findings are particularly
important to consider, given that we can not ethically justify replicating
these findings without eliminating the now known risks to participants.
We urge any other researchers with relevant data to come forward, for
the research community should not quietly conclude that bolstering
self–esteem harms performance while thousands of schools continue to
bolster self–esteem in the misguided but unquestioned belief that it will
Limitations of this study must be acknowledged. We studied a rela-
tively small number of students (although the effect was sufficiently ro-
bust to reach significance). The self–esteem bolstering was well inten-
tioned and resembles other interventions in content, but it is in principle
conceivable that other interventions would have the desired (opposite)
effect. Our self–esteem intervention was also designed to bolster both
general and academic self–esteem, and so the findings do not inform
multidimensional models of the self (e.g., Marsh & Craven, 1997). The
use of email, too, may have also produced different effects from
456 FORSYTH ET AL.
face–to–face self–esteem bolstering. Generalizing across cultural and
generational boundaries is hazardous. Twenge’s (2006) cross–temporal
meta–analyses have shown that the current generation of young adults,
who comprised our sample, differs substantially and significantly from
previous generations in several relevant respects. Most notably, they
have higher self–esteem and lower beliefs in internal control. The pres-
ent manipulations of self–esteem and responsibility attribution thus
were delivered to members of a generation who are already relatively
high on one dimension and low on the other.
Also, all members of this sample had just received a grade of C or be-
low on the first test in an introductory psychology course. It is entirely
possible that bolstering self–esteem among high–achieving students
would be free of ill effects and might even be beneficial. Some research-
ers have speculated that boosting self–esteem in recognition of success-
ful achievement may be an effective strategy to promote good perfor-
mance and socially desirable outcomes (e.g., Baumeister et al., 2003).
The present findings indicate merely that boosting self–esteem or en-
couraging feelings of controllability in the wake of poor performance
can be costly and counterproductive.
The self–esteem movement that began in the 1970s appears to have
succeeded, at least to the degree that it raised self–esteem scores across
North America. A meta–analysis by Twenge and Campbell (2001) indi-
cated that the median male college student around 1995 had higher
self–esteem than 86% of his peers in 1968 had. Another meta–analysis by
Haney and Durlak (1998) confirmed that the self–esteem boosting
school programs do succeed, on average, at increasing scores on self–es-
teem measures. (Whether this reflects actual change in self–concept, or
merely enhanced willingness to rate oneself positively on question-
naires, remains unclear, even in the national data.) Narcissism, mean-
while, may have increased even more than self–esteem (Twenge, 2006).
Despite these well documented improvements in self–appraisal, SAT
scores have gone down over the same period. That seemingly paradoxi-
cal pattern fits well with the present findings, which suggest that bol-
stering self–appraisals in the absence of objective or contingent success
will make school performance worse rather than better.
In sum, we continue to believe that adequate self–esteem and a sense
of control are hallmarks of good adjustment (see also, Taylor & Brown,
1988). The present findings suggest only that bolstering them directly,
independent of actual performance, may be ineffective as a strategy for
improving students’ work (at least for students whose grades are al-
ready below average). Possibly self–esteem boosts can help learning by
serving as a reward for good performance. Perhaps self–esteem boosts
that stress hard work, strategizing, and persistence can enhance perfor-
BOLSTERING SELF–ESTEEM 457
mance without the accompanying pitfalls of self–praise unrelated to
performance (Kamins & Dweck, 1999). But persuading students to think
well of themselves despite having performed poorly on a first test
seems, if anything, to make students do even worse.
Bachman, J. G., & O’Malley, P. M. (1977). Self–esteem in young men: A longitudinal analy-
sis of the impact of educational and occupational attainment. Journal of Personality
and Social Psychology, 35, 365–380.
Bachman, J. G., & O’Malley, P. M. (1986). Self–concepts, self–esteem, and educational expe-
riences: The frog pond revisited (again). Journal of Personality and Social Psychology,
Baumeister, R. F. (1998). The self. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook
of social psychology (4th ed., Vol. 1, pp. 680–740). New York: McGraw–Hill.
Baumeister, R. F., Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003). Does high self–esteem
cause better performance, interpersonal success, happiness, or healthier lifestyles?
Psychological Science in the Public Interest,4, 1–44.
Crocker, J., & Major, B. (1989). Social stigma and self–esteem: The self–protective proper-
ties of stigma. Psychological Review, 96, 608–630.
Crocker, J., & Park, L. E. (2004). The costly pursuit of self–esteem. PsychologicalBulletin,130,
Diener, C. I., & Dweck, C. S. (1980). An analysis of learned helplessness II: The processing
of success. Journal of Personality and Social Psychology,39, 940–952.
Di Paula, A., & Campbell, J. D. (2002). Self–esteem and persistence in the face of failure.
Journal of Personality and Social Psychology, 83, 711–724.
Dunning, D., Heath, C., & Suls, J. M. (2004). Flawed self–assessment: Implications for
health, education, and the workplace. Psychological Science in the Public Interest, 5,
Dweck, C. S. (1975). The role of expectations and attributions in the alleviation of learned
helplessness. Journal of Personality and Social Psychology,31, 674–685.
Forsyth, D. R. (1986). An attributional analysis of students’ reactions to success and failure.
In R. Feldman (Ed.), The social psychology of education (pp. 17–38). New York: Cam-
bridge University Press.
Haney, P., & Durlak, J. A. (1998). Changing self–esteem in children and adolescents: A
meta–analytic review. Journal of Clinical Child Psychology, 27, 423–433.
Kamins, M. L., & Dweck, C. S. (1999). Person versus process praise and criticism: Implica-
tions for contingent self–worth and coping. Developmental Psychology, 35, 835–847.
Leary, M. R. (2004). The curse of the self: Self–awareness, egotism, and the quality of human life.
New York: Oxford.
Lodewyk, K. R., & Winne, P. H. (2005). Relations among the structure of learning tasks,
achievement, and changes in self–efficacy in secondary students. Journal of Educa-
tional Psychology, 97, 3–12.
Maruyama, G. M., Rubin, R. A., & Kingsbury, G. G. (1981). Self–esteem and educational
achievement: Independent constructs with a common cause? Journal of Personality
and Social Psychology, 40, 962–975.
McFarlin, D. B., & Baumeister, R. F. (1984). On knowing when to quit: Task failure, self–es-
teem, advice, and nonproductive persistence. Journal of Personality, 52, 138–155.
458 FORSYTH ET AL.
Marsh, H. W., & Craven, R. (1997). Academic self–concept: Beyond the dustbowl. In G.
Phye (Ed.), Handbook of classroom assessment: Learning, achievement, and adjustment
(pp. 131–198). Orlando, FL: Academic Press.
Noel, J., Forsyth, D. R., & Kelley, K. (1987). Improving the performance of failing students
by overcoming their self–serving attributional biases. Basic and Applied Social Psy-
chology, 8, 151–162.
Scheirer, M. A., & Kraut, R. E. (1979). Increasing educational achievement via self–concept
change. Review of Educational Research, 49, 131–150.
Seligman, M. E. P. (1975). Helplessness: On depression, development and death. San Francisco,
Sommers, C. H. (1995). Who stole feminism? How women have betrayed women. New York: Si-
mon & Schuster.
Sternberg, R. J. (1998). Principles of teaching for successful intelligence. Educational Psy-
chologist, 33, 65–72.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well–being: A social psychological perspec-
tive on health. Psychological Bulletin, 103, 193–210.
Twenge, J. M. (2006). Generation Me. New York: Free Press.
Twenge, J. M., & Campbell, W. K. (2001). Age and birth cohort differences in self–esteem: A
cross–temporalmeta–analysis. Personality andSocial PsychologyReview, 5, 321–344.
Twenge, J. M., & Crocker, J. (2002). Race and self–esteem: Meta–analyses comparing
Whites,Blacks, Hispanics,Asians, and American Indians. Psychological Bulletin, 128,
Valentine, J. C., DuBois, D. L., & Cooper, H. (2004). The relation between self–beliefs and
academic achievement: A meta–analytic review. Educational Psychologist, 39,
Wilson, T. D., & Linville, P. W. (1982). Improving the academic performance of college
freshmen: Attribution therapy revisited. Journal of Personality and Social Psychology,
Wilson, T. D., & Linville, P. W. (1985). Improving the performance of college freshmen with
attributional techniques. Journal of Personality and Social Psychology,49, 287–293.
Wylie, R. C. (1979). The self–concept: Vol. 2. Theory and research on selected topics. Lincoln: Uni-
versity of Nebraska Press.
BOLSTERING SELF–ESTEEM 459