You are on page 1of 17

Global Self-Esteem and Specific Self-Esteem: Different Concepts, Different Outcomes

Author(s): Morris Rosenberg, Carmi Schooler, Carrie Schoenbach and Florence Rosenberg
Source: American Sociological Review, Vol. 60, No. 1 (Feb., 1995), pp. 141-156
Published by: American Sociological Association
Stable URL: http://www.jstor.org/stable/2096350
Accessed: 08-05-2015 08:44 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to American Sociological
Review.

http://www.jstor.org

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM:
DIFFERENT CONCEPTS, DIFFERENT OUTCOMES

Morris Rosenberg Carmi Schooler


University of Maryland and National Institute of Mental Health
National Institute of Mental Health

Carrie Schoenbach Florence Rosenberg


National Institute of Mental Health Walter Reed Army Institute of
Research and UniformedServices
University of the Health Sciences

In this paper, we attempt to shed light on the nature of relevance of, and
relationship between global self-esteem and specific self-esteem. We mar-
shal evidence that the two types of self-esteem may have strikingly different
consequences, global self-esteem being more relevantto psychological well-
being, and specific self-esteem being more relevant to behavior. We use lin-
ear structural equation causal modeling to test this hypothesisfor the case
of global self-esteem (Rosenberg 1979) and specific (academic) self-esteem.
Ourfindings show that, while global self-esteem is more strongly related to
measures of psychological well-being, specific (academic) self-esteem is a
much better predictor of school performance. Otherfindings indicate that
the degree to which specific academic self-esteem affects global self-esteem,
particularly the positive component of global self-esteem, is a-function of
how highly academic performance is personally valued.

O )ver the past 40 years, the concept of Looking at the general body of research on
self-esteem has assumed an important self-esteem today, it is evident that most of
place in the field of social psychology. A this literature deals with global self-esteem,
computer search of the literature (Kitano that is, the individual's positive or negative
1989) found over 6,500 article titles that ex- attitude toward the self as a totality. In the
plicitly used the term "self-esteem" and over last decade, however, a number of writers
30,000 titles that used the term "self' in have stressed the importanceof studying spe-
some hyphenated form, many of which also cific self-esteem, as well (e.g., Harter 1985;
dealt with self-esteem (e.g., self-concept, Marsh 1986; Marsh and Shavelson 1985;
self-evaluation, self-respect, self-confi- Swann 1987). As Marsh (1990) expresses it:
dence). "More recently, self-concept theory has
stressed the multi-dimensionality of self-
*Direct all correspondence to Carmi Schooler, concept, and empirical studies have identi-
Laboratory of Socio-environmental Studies,
fied distinct, a priori facets of self-concept"
NIMH, NIH, Rm BIA-14, Federal Building, 7550
Wisconsin Avenue, Bethesda, MD 20892. The (p. 107).
authors presented an earlier version of this paper The aim of this paper is to shed light on
to Research Committee 42, Social Psychology, at the nature and relevance of global and spe-
the 12th World Congress of Sociology, Madrid, cific self-esteem and their relationship to
Spain, July 1990. This paper, which represents a each other. We begin by focusing on two
very substantial revision of the earlier one and general features of attitudes to clarify the
includes new analyses, was completed after distinctions we make regarding self-esteem.
Morris Rosenberg's untimely death in 1992. First, the study of any attitude, and self-es-
Among the analyses completed after Rosenberg's
teem is an attitude, must take account of the
death are those that separatethe positive and nega-
tive aspects of self-esteem and those that examine fact that people may
have attitudes both to-
how valuing academic performancerelates to spe- ward an object as a whole (global or general)
cific academic self-esteem and global self-esteem. and toward specific "facets" of that object

American Sociological Review, 1995, Vol. 60 (February:141-156) 141

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
142 AMERICAN SOCIOLOGICAL REVIEW

(Marsh 1990). For example, a student may that "negative self-conceptions may help
have attitudes toward her university as a people maintaina viable self-system and pre-
whole, but she may also have different atti- dictable orderly social relations"(p. 289). On
tudes toward a specific department,the qual- the other hand, "'effectance motivation' that
ity of the faculty, or the attractivenessof the compels judgment of one's own competence
campus. Although the differences between and efficacy (Gecas and Schwalbe 1983) ...
global and specific attitudes are sometimes should impel one to focus more sharply on
overlooked, they are not equivalent or inter- one's own varying degrees of ability, compe-
changeable. This point applies equally to tence, and efficacy-all attributes of self-
self-esteem, which can be viewed as an atti- confidence" (Owens 1993:290).
tude toward an object, even though the On a more empirical level, Owens (1993)
holder of the attitude and the object toward has reconfirmed the existence of the self-
which the attitude is held-the self-are the confidence and self-deprecation components
same (Rosenberg 1979). of self-esteem. He also has shown that both
A second feature of attitudes is that they components fit well into a second-order
include both cognitive and affective ele- construct of global self-esteem, with self-
ments. That attitudes are cognitive is evident confidence having an appreciably higher as-
from the fact that they refer to objects-an sociation than self-deprecation with global
attitude represents some thought about a par- self-esteem (e.g., at age 23 the effect of glo-
ticular thing (e.g., person, material object, bal self-esteem on self-confidence is .88 and
group, idea, etc.). That they are affective is on self-deprecation is -.64). We will con-
shown by the fact that attitudes have both di- sider this issue further in our analyses sec-
rection (i.e., a positive or negative orienta- tion.
tion toward some object) and intensity. Self- We have defined the following objectives:
esteem research has tended to overlook the
(1) To illustrate how investigators have
degree to which these cognitive and affective
sometimes reached incorrect conclu-
components differentially permeate specific
and global self-esteem. sions by treatingglobal and specific self-
esteem as surrogatesfor one another.
Like other attitudes, individuals' views of
themselves can contain both positive and (2) To demonstrate that global and specific
negative components. Thus, decades ago or- self-esteem differ in their consequences
thogonal principal component factor analy- and how these different consequences
ses carried out by Kohn and Schooler (1969) have led to some of the mispredictions
indicated that the Rosenberg Self-Esteem and weak associations reportedin the lit-
Scale (Rosenberg 1979) contained two com- erature.
ponents-self-confidence and self-depreca-
(3) To show that the effects of global and
tion. Using structural equation based mea-
specific self-esteem may be mediated by
surement models they later confirmed that a
each other.
two-component model that separates the
positive and negative aspects of self-esteem (4) To examine how global and specific self-
provides a better fit to the data than does the esteem affect one another, with particu-
single general component model that charac- lar emphasis on whether the effect of a
terized Rosenberg's (1979) original concep- specific facet of self-esteem on global
tion (Kohn and Schooler 1983). self-esteem is a function of the degree to
More recently, Owens (1993) theorized which that facet is valued.
that, in addition to self-esteem and self-con-
(5) To explore how global and specific self-
sistency, the two central motives credited by
esteem differentially affect, and are dif-
self-concept theory for protecting and main-
ferentially affected by, behavioral out-
taining one's self-image, other theoretical
comes.
perspectives suggest plausible functions for
the less global concepts of self-deprecation (6) To determine the relative importance of
and self-confidence. According to Owens, the positive and negative components of
for example, self-verification theory (Swann, global self-esteem with respect to a be-
Stein-Seroussi, and Giesler 1992) suggests havioral outcome, school performance.

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 143

THE FALLACYOF EQUATING GLOBAL of comparison groups but also on ignoring or


AND SPECIFIC SELF-ESTEEM devaluing problematic areas of functioning
(Pearlin and Schooler 1978). All these possi-
In more than a few self-esteem studies, the bilities are consistent with the hypothesis
failure to distinguish the parts from the that decreasing the value one gives to do-
whole has led to a numberof misunderstand- mains in which oneself or one's reference
ings. A well-known example is the large group does poorly protects global self-es-
body of research on the relationship between teem (Harter 1985; Rosenberg 1979). If this
race and self-esteem. Some studies focus on hypothesis is correct, global and specific
pride in one's race (racial self-esteem), self-esteem are different phenomena-they
whereas others focus on pride in oneself may be dynamically interrelatedphenomena,
(personal self-esteem) (Porter and Washing- but they are not directly interchangeable.
ton 1989). This distinction is often over- Marsh's (1986) research provides clear
looked. In his exhaustive review of the lit- empirical evidence that global and specific
erature,Cross (1985) pointed out that 87 per- self-esteem cannot serve as surrogates for
cent of the 161 studies he examined explored one another.He examined the relationship of
either racial self-esteem or personal self-es- subjects' self-evaluations on 12 facets of the
teem, but not both. In most studies, low re- self to their global self-esteem. He found that
gard for one's race was considered equiva- the associations ranged from .06 to .60, with
lent to low regard for oneself. Cross (1985) most falling within the .30 to .50 range. Even
reportedthat, although researchersclearly set if we accept the possibility that measurement
out to assess racial self-esteem, as a result of errormight lower the correlations somewhat,
their failure to make a sharp distinction be- such relatively low correlations are exactly
tween racial and personal self-esteem, they what one would expect if specific and global
often presented the racial self-esteem results self-esteem are related but not interchange-
"as if' their study had actually assessed level able phenomena.
of self-worth. Similarly, children with poor It might seem apparent that the relation-
academic self-concepts are often described ship between one's judgment of a particular
as having low self-esteem, based on the im- facet of oneself and one's global self-esteem
plicit assumption that a child who feels that would depend in large part on the rank of
he or she is a poor student is thereby express- that facet in one's personal hierarchy of self-
ing low general self-esteem (Wylie 1979). values. Although this point underlies the
Clearly, the aspects of oneself on which point made in the preceding paragraph and
one's global self-esteem is based and the was clearly noted many years ago by Will-
choice of reference group to which one com- iam James ([1890] 1950), not all empirical
pares oneself are complex matters. There is findings unequivocally supportthe relevance
considerable experimental evidence (Major, of psychological centrality. (For examples of
Sciacchitano, and Crocker 1993) and survey- mixed evidence, see Faunce 1982; Gecas and
based evidence (Rosenberg 1979) supporting Seff 1990; Harter 1985; Hoelter 1986; Hoge
the hypothesis that members of stigmatized and McCarthy 1984; Marsh 1986; Rosenberg
groups avoid threats to their self-esteem by 1989; Thomas 1989.) Using data measuring
comparing themselves primarily with others how much our respondents value academic
who are members of their own stigmatized achievement, we empirically test whether the
group rather than with members of an ad- relationship between global self-esteem and
vantaged out-group (Crocker and Major specific academic self-esteem varies when
1989). Other evidence indicates that mem- the value placed on academic achievement
bers of stigmatized or disadvantaged groups varies.
also protect their self-esteem by selectively
devaluing those domains in which the out-
RELATIONSHIPS OF GLOBAL AND
group is advantaged and selectively valuing
SPECIFIC SELF-ESTEEM TO
those domains in which their in-group has
BEHAVIOR AND WELL-BEING
advantages (Major et al. 1993). Research on
coping suggests that other general coping It is particularlyimportantto distinguish be-
strategies exist, based not only on the choice tween global and specific self-esteem be-

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
144 AMERICAN SOCIOLOGICAL REVIEW

cause the relationships reportedin the litera- ger relationships between specific self-es-
ture between self-esteem and other variables teem and a behavioral outcome. As noted,
are often weaker than might be expected. when we speak of specific self-esteem, we
This is true whether self-esteem is treated as refer to a particularfacet of the self (Marsh
cause or outcome. For example, Wylie 1986). Insofar as this facet relates to some
(1979) and others reported that sociodemo- area of competence, specific self-esteem has
graphic variables show no better than mod- much in common with the concept of self-
est success in predicting self-esteem. (Some efficacy. When Bandura (1982) speaks of
reasons for these results have been discussed self-efficacy, he refers to the individual's con-
by McCarthy and Yancey 1971; Rosenberg fidence that he or she can attainspecified per-
and Pearlin 1978; Rosenberg and Simmons formance levels. Bandura (1982) identified
1972). Although there have been some suc- several reasons why perceived self-efficacy
cesses (e.g., Menaghan and Parcel 1990), in tends to enhance performanceoutcomes. One
general, self-esteem has not proved to be an reason, he noted, is that "people who judge
impressive predictorof behavioraloutcomes. themselves ineffective in coping with envi-
One reason for these comparatively weak ronmental demands tend to generate high
relationships between self-esteem and behav- emotional arousal, become excessively pre-
ior has been the failure to recognize that glo- occupied with personal deficiencies, and
bal and specific self-esteem are both impor- cognize potential difficulties as more formi-
tant, but that they are importantfor different dable than they really are. Such self-referent
reasons and are relevant in different ways. A concerns undermineeffective use of the com-
central hypothesis of this paper is that spe- petencies people possess" (pp. 25-26).
cific self-esteem is most relevant to behav- Another reason Bandura(1982) offered for
ior, whereas global self-esteem is most rel- perceived self-efficacy to result in successful
evant to psychological well-being. Much performanceis that "self-judged efficacy ...
self-esteem research, we believe, relates pre- determines how much effort people will ex-
cisely the wrong type of self-esteem to the pend and how long they will persist in the
outcome variable (Scheff, Retzinger, and face of obstacles and aversive experi-
Ryan 1989), examining the relationship be- ences.... In the face of difficulties, people
tween global self-esteem and specific behav- who entertain serious doubts about their ca-
iors or behavioral outcomes. We believe, on pacities slacken their efforts or give up alto-
the contrary, that a specific behavior is best gether, whereas those who have a strong
predicted by a specific self-esteem that is in sense of efficacy exert greater effort to mas-
some way connected to that behavior, ter the challenges" (p. 25). At comparable
whereas psychological well-being is best ability levels, then, self-efficacious persons
predicted by global self-esteem. Let us con- are more apt to experience successful out-
sider some of the evidence on which our hy- comes.
pothesis is based. Global self-esteem, on the other hand,
seems much less likely to exercise a power-
ful direct effect on performance. For one
Self-Esteem and Behavior thing, the central feature of global self-es-
The expectation that specific. self-esteem teem appears to be self-acceptance or self-
would have stronger effects on behavior than respect. Competence is only one factor, and
global self-esteem derives from the Fishbein not necessarily the most importantone, con-
and Azjen (1975) model, which postulates tributing to such feelings. Second, as we
that the power of an attitude to predict a be- have noted, some facets of the self may be
havior is a function of how closely that atti- peripheralto feelings of self-worth, whereas
tude relates to the act in question-the more others may be central. Unless a particular
specific the attitude, the greaterits predictive facet is important to the individual, there is
power. If so, then a specific self-esteem little reason to think that global self-esteem
should be a better predictor of a specific be- will tell us much about a person's behavior
havior than is global self-esteem. or performancewith respect to that facet, nor
Self-efficacy theory and self-attribution will such behavior necessarily indicate much
theoryoffer other reasons for expecting stron- about global self-esteem. For example,

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 145

knowing that someone has high global self- reason to believe that it is a decidedly better
esteem will tell us little about that person's predictor of psychological well-being.
assessment of his or her competence as a The theoretical foundation for this expec-
pole-vaulter or as a writer of sonnets. Con- tation lies in "self-enhancement theory"
versely, the fact that I consider myself totally (Baumeister 1982; Greenwald 1980; Jones
inept as a pole-vaulter in itself offers little 1973; Kaplan 1975; Swann 1987), which
indication of my overall feeling of self- states that self-esteem is a fundamental hu-
worth. We do not suggest that global self-es- man motive. Thus, the self-esteem motive
teem is totally unrelatedto behavior, but that (also called the "self-maintenance motive"
this relationship is likely to be weaker than by Tesser and Campbell [1983] and the "mo-
the relationship of specific self-esteem to a tive for self-worth"by Covington [1984]) has
relevant behavior or performance. been identified by Maslow (1970) as one of
Although evidence is limited, the results of the "prepotent"human needs. All of these
both experimental and correlational studies theories share the view that there exists in
support this view. In an experimental study, human beings a universal desire to protect
Shrauger (1972) asked female undergradu- and enhance their feelings of self-worth and
ates to specify what percentage of under- that the frustration of this desire generates
graduates would perform better than them- some measure of psychological distress.
selves on a "concept attainment task." The Maintenanceof self-esteem leads to self-pro-
percentage score served as the measure of tective motives, self-enhancement processes,
specific self-esteem. General (global) self- and a variety of coping processes. As noted
esteem was based on a modification of a self- earlier in the discussion of the fallacy of
esteem measure used extensively by Diggory equating global with specific self-esteem,
(1966). The concept attainment task was a such self-maintenance motives and processes
modification of a measure developed by have been shown to affect the comparison
Weick (1964). Shrauger found that specific group one chooses as well as how one reacts
self-esteem was a significant predictorof ac- to unfavorable comparisons (Major et al.
tual performance on the concept attainment 1993).
test, but that global self-esteem was not. That global self-esteem is associated with
Correlational findings are consistent with psychological well-being has been demon-
the experimental results. In Bachman's strated repeatedly in past research. One
(1970) study of tenth-gradeboys, the correla- firmly established finding in this literatureis
tion between global self-esteem and school the inverse association between self-esteem
marks was .23, whereas the correlation be- and depression (Rosenberg 1985; Wylie
tween school marks and self-concept of aca- 1979). Studies of children, adolescents,
demic ability was .48. Reviewing a number adults, and the elderly all show the same pat-
of studies dealing with the relationship be- tern. Pearlin and Lieberman's (1979) study
tween self-esteem and academic perfor- of 2,300 adults in the Chicago Metropolitan
mance, Wylie (1979) concluded that the cor- Area showed a correlation of -.49 between
relation between global self-esteem and grade self-esteem and depression. Similar findings
point average is usually about .30, whereas appear in Kaplan and Pokorny (1969) and
the association between specific self-esteem Rosenberg and Simmons (1972). Using struc-
(academic self-concept) and grade point av- tural equation models to examine the causal
erage is more likely to be in the range of .45 relationships underlying these frequently re-
to .70. Wylie (1979:698-700) also reporteda ported correlations, we have found that self-
number of other instances in which specific esteem and depression significantly affect
self-esteem is correlatedwith specific behav- each other, although the negative relationship
ior but global self-esteem is not. between the two variables seems to be due
somewhat more to the effect of depression on
self-esteem than to the effect of self-esteem
Self-Esteem and Psychological Well-Being on depression (Rosenberg, Schooler, and
Although global self-esteem is less likely Schoenbach 1989). Clinical evidence (e.g.,
than specific self-esteem to be a good pre- Beck, Rush, Shaw, and Emery 1979) supports
dictor of behavior or performance, there is these quantitativeresults. Global self-esteem

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
146 AMERICAN SOCIOLOGICAL REVIEW

has also been shown to be strongly related to vestigation: depression (6 items), anomie (8
levels of anxiety, whether expressed in so- items), general anxiety (7 items), resent-
matic symptoms (Luck and Heiss 1972; ment (7 items), anxiety and tension (5
Rosenberg 1989) or in psychological mani- items), irritability (7 items), life satisfaction
festations (Luck and Heiss 1972; Rosenberg (1 item), guilt (5 items), happiness (6
1985). Specific self-esteem, on the other items), and negative affective states (4
hand, may have little direct effect on psycho- items). In Tables 1 and 2, the data for these
logical well-being, and, as noted above, what measures are from summary scores on the
effects it does have may be influenced by the items.2
psychological centrality of the particular
facet of the self that is involved.
RESULTS
To summarize our hypotheses, with respect
to behavior or behavioral outcomes, we ex- Table 1 presents the zero-order correlation
pect specific self-esteem to be a better pre- coefficients for each of these 10 well-being
dictor than global self-esteem. On the other indices with global self-esteem on the one
hand, with regard to psychological well-be- hand, and with specific (academic) self-es-
ing, global self-esteem will be a better pre- teem on the other.3
dictor than specific self-esteem. No previous Table 1 clearly shows that global self-es-
studies, as far as we know, have made these teem is more strongly related to most mea-
comparisons. sures of psychological well-being (depres-
sion, anomie, general anxiety, resentment,
anxiety-tension, irritability, life satisfaction,
DATA AND MEASURES happiness, and negative affective states)
We use data drawn from the Youth in Transi- than is specific self-esteem. The mean
tion study (Bachman 1970). This longitudi- strength of the relationship of global self-es-
nal study, based on a probability sample of teem to all 10 measures is .337, whereas the
2,213 tenth-grade boys in 87 high schools mean strength relationship of academic self-
throughoutthe contiguous 48 states, includes esteem to these measures is .079. The single
four waves extending over an eight-year exception is the guilt variable, which is un-
span. Our analysis is limited to the 1,886 related to global self-esteem but is signifi-
boys who participated in the first two waves cantly related to academic self-esteem. On
of the study-1966 and 1968. the -other hand, when we consider our spe-
In our analysis, global self-esteem is mea- cific behavioral outcome (school perfor-
sured by 6 of the 10 items of the Rosenberg mance), then specific (academic) self-es-
Self-Esteem Scale (Rosenberg 1989). Aca- teem turns out to be much more highly cor-
demic self-esteem (our specific self-esteem related than is global self-esteem (.488 ver-
measure) is based on a 3-item index.' sus .253). This pattern of correlations indi-
School marks, which serve as the behavioral cates that self-esteem is significantly related
outcome, are based on self reported grade to a number of other variables, but only if
point average. The following psychological
well-being measures are included in this in- 2These indices and the items used to create
them are described in the Documentation Manual
I We use Bachman's School Ability Self-Con- for Bachman's Youth in Transition study (Bach-
cept Index, which is based on subjects' responses man 1975).
to three questions. Two questions are coded on a 3 We measure global self-esteem using an in-
scale from 1 through 6-far below average to far dex created by summing the responses to six
above average: (1) "How do you rate yourself in Rosenberg Self-Esteem Scale items from the
school ability compared with those in your grade Bachman (1970) data (coded from 1 through 5-
in school?"; (2) "How intelligent do you think never true to almost always true): (1) "I feel I'm
you are compared to others your age?" The third a person of worth"; (2) "I feel I have a number of
question, (3) "Comparedto others your age, how good qualities"; (3) "I am able to do things as
important is it to you to be able to use your intel- well as most other people"; (4) "I feel I do not
ligence?" is coded from 1 through 5-much less have much to be proud of"; (5) "I take a positive
important than average to much more important attitude toward myself"; (6) "Sometimes I think I
than average. am no good at all."

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 147

Table 1. Zero-Order Correlation Coefficients of Global and Specific Mediators


Psychological Well-Being and Behav-
ioral Outcome with Global Self-Esteem Although the data in Table 1 are generally
and Specific (Academic) Self-Esteem: consistent with our hypothesis that global
Tenth-Grade Boys, 1966 and 1968
self-esteem is a better predictor of psycho-
Global Academic logical well-being, whereas specific self-es-
Outcome Variable Self-Esteem Self-Esteem teem is a better predictor of behavioral out-
comes, it is nevertheless true that some dis-
Psychological Well-Being crepant effects appear.For example, there is
Depression -.432*** -.103*** a significant association between global self-
Anomie -.409*** -. 154*** esteem and school marks, and there are some
Generalanxiety -.232*** -.010 small but significant associations between
Resentment -.346*** -.120*** specific self-esteem and certain measures of
psychological well-being. To a modest ex-
Anxiety/tension -.319*** -.023
tent, then, global self-esteem does predict
Irritability -.353*** -.044 behavior and specific self-esteem does pre-
Life satisfaction .318*** .052* dict psychological well-being. Do these find-
Guilt -.033 .087*** ings represent exceptions to our generaliza-
Happiness .499*** .105***
tion? Not necessarily. Global and specific
self-esteem may each mediate the effect of
Negative affective -.424*** -.093*** the other. For example, if global self-esteem
states
is associated with school marks because glo-
Behavioral Outcome bal self-esteem affects academic (specific)
School marks(GPA) .253*** .488*** self-esteem and academic self-esteem in turn
affects school marks, then it would still be
p <.05 **p< .01 ***p< .001 (two-tailedtests)
true that specific self-esteem is directly re-
Note: N = 1,886. sponsible for the behavioral outcome. Glo-
bal self-esteem would "depend on" specific
the appropriate type of self-esteem (global self-esteem to exercise its behavioral effect.
or specific) is paired with the appropriate Conversely, if the association between spe-
correlate (psychological well-being or per- cific self-esteem and psychological well-be-
formance). ing were attributableto the fact that specific
These data also tell us something about the self-esteem affected global self-esteem and
nature of both types of self-esteem. Atti- that global self-esteem, in turn, affected psy-
tudes, as we pointed out earlier, have both chological well-being, then one could still
cognitive and affective elements. This is as say that global self-esteem is directly respon-
true of attitudes toward the self as toward sible for psychological well-being.
anything else, but the affective and cognitive Table 2 presents the coefficients from the
components of self-attitudes may not be multiple regression of both types of self-es-
equally representedin global self-esteem and teem (global and specific) on both types of
specific self-esteem. Given the stronger as- outcomes (psychological well-being and
sociation of global self-esteem with psycho- school marks). It is difficult to test the hy-
logical well-being, it is probable that global pothesis that the relationship between spe-
self-esteem is chiefly an expression of per- cific self-esteem and psychological well-be-
sonal affect. And given the stronger associa- ing is largely mediated by global self-esteem,
tion of specific self-esteem with behavioral because the original relationship is so weak,
outcome, it is likely that specific self-esteem, althougheven this weak relationshipbecomes
which is probably largely a matter of judg- weaker still when global self-esteem is con-
ment or evaluation of a particularcharacter- trolled (average correlation declines from
istic, may be more cognitive in nature. It is .079 to .067). Controlling academic self-es-
thus understandablethat these two types of teem also does not diminish the relationship
self-esteem should have different effects- of global self-esteem to psychological well-
each effect, of course, being very important being (average correlation actually increases
in its own right. from .337 to .359). When we turn to the be-

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
148 AMERICAN SOCIOLOGICAL REVIEW

Table 2. Coefficients from Multiple Regressions equally plausible theoretical arguments that
of Psychological Well-Being and Behav- the specific causes the global or that the glo-
ioral Outcome on Global Self-Esteem bal causes the specific: Thus, since global
and Specific (Academic) Self-Esteem:
self-esteem is in some sense based on the
Tenth-Grade Boys, 1966 and 1968
judgments of various parts of the self, the
IndependentVariable parts (specifics) might be seen as responsible
for the whole (global). On the other hand,
Global Academic
DependentVariable Self-Esteem Self-Esteem
assessments of particular facets of the self
may well be based on one's overall feelings
Psychological Well-Being of self-worth. It is therefore requisite to em-
Depression -.451*** .056* pirically examine whether the parts are
Anomie -.405*** -.011 chiefly responsible for the whole or the
Generalanxiety -.261 ***
whole for the parts.
.082***
This question has rarely been investigated
Resentment -.346*** .002
systematically in the literature. Perhaps this
Anxiety-tension -.355*** .101*** history of neglect is attributableto the lack
Irritability -.385*** .092*** of a satisfactory methodology for dealing
Life satisfaction .342*** -.069** with possible reciprocal effects and to the
Guilt -.072** .113*** stringent data requirementsfor the appropri-
Happiness .527*** -.081 *** ate analyses. Today, however, the availabil-
Negative affective -.447*** .065** ity of structural equation modeling proce-
states dures (Joreskog and Sorbom 1976) and panel
data sets make it possible to specify the ef-
Behavioral Outcome fect each variable has on the other.
School marks(GPA) .093*** .455*** We begin by examining the reciprocal ef-
*p<.05 ** <.01 p < .001 (two-tailedtests) fects of academic self-esteem and global
Note: N = 1,886. This table reportsthe results of self-esteem.4 Several features of the causal
11 multipleregressionanalyses, each of which used aspects of these models should be noted.
global self-esteem and academic self-esteem as the First, we use panel data (waves 1 and 2,
two independentvariables. based on 1,886 respondents), which helps to
solve the structural equation models and to
estimate the causal influences. To solve the
havioral outcome, however, the data suggest simultaneous equations necessary for esti-
that global self-esteem affects specific (aca- mating reciprocal effects, instrumentationis
demic) self-esteem, which, in turn, affects necessary: An instrument is a variable that
school marks.We see that, when specific self- identifies an equation by not being allowed
esteem is held constant through multiple re- to have a direct effect on the dependent vari-
gression, the relationship between global able of that equation. For an instrument to
self-esteem and school marksis substantially identify an equation meaningfully, there
reduced (from .25 to .09). On the other hand, must be good theoreticaljustification that its
the relationship between specific self-esteem
and school marks is almost unaffected when 4 The measure of global self-esteem used in
global self-esteem is controlled (change from Tables 1 and 2 is based on summary scores. In
.49 to .46). These findings, then, are consis- the present analysis, global self-esteem (based on
tent with the view that specific self-esteem the same items) is estimated for 1966 and 1968
has a direct effect on behavior (or behavioral as part of a full-information linear structural
outcomes), whereas global self-esteem has a equation model (Joreskog and Sorbom 1976). All
direct effect on psychological well-being. of the correlations of residuals of the same vari-
ables at both points in time are estimated. In ad-
dition, significant correlations of residuals be-
Global and Specific Self-Esteem: Which tween different variables are included where sug-
Causes Which? gested by an examination of the first-order par-
tial derivatives. In all structural equations dis-
What is the causal connection between glo- cussed, we allow errors in the causal equations to
bal and specific self-esteem? One can make be correlated.

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 149

effect should only be indirect, and the instru- more general. This is why Marsh and Shavel-
ment must be reasonably correlated with the son conceptualized the self-concept as a hi-
variable it is not allowed to affect directly. In erarchical structure. Consider, for example,
our model, cross-lagged effects are omitted facets of the self-concept that exist in the in-
to provide a source of instrumentationfor the tellectual realm. At a highly specific level
reciprocal effects that are modeled for wave might be a child's assessment of his or her
2. These wave 2 effects are viewed as out- reading ability or math ability (Marsh 1986).
comes of long-term processes and thus rep- At a somewhat broader level would be the
resent the sum of the lagged and contempo- child's academic self-esteem, which would
raneous effects. be based in part on math and reading ability
Additional instrumentalvariables are also and in part on other facets. Still broader
employed. Instrumentsfor the global to aca- would be the child's assessment of his or her
demic self-esteem relationship are positive general level of intelligence, which would be
family relationships, numberof best friends, affected by academic self-esteem but also by
physical appearance,and complexion. All of other facets. When we consider the recipro-
these instrumental variables could be ex- cal effects of global and specific self-esteem,
pected to directly affect global self-esteem. then, it is importantto consider the level of
None of them would be expected to have a specificity or generality of the specific self-
direct effect on academic self-esteem; any concept facet.
effect would be indirect through global self- We addressed this question by examining
esteem's effect on academic self-esteem. In- the reciprocal effects of global self-esteem
strumentsfor the academic to global self-es- and the students' estimates of their own in-
teem relationship are hours spent on home- telligence. Self-estimate of intelligence was
work, positive school attitudes, negative measuredby responses to the following item:
school attitudes, and valuation of academic "How intelligent do you think you are com-
achievement. All of these instrumental vari- pared with other boys your age?" Six catego-
ables would be expected to affect academic ries of response were provided ranging from
self-esteem without directly affecting global "Far below average (bottom 10 percent)" to
self-esteem. / "Far above average (top 10 percent)." Note
We introduce the following variables as that this item was also included in the aca-
controls affecting both global self-esteem demic self-esteem scale; however, compared
and academic self-esteem: race, age, intact to the other items in that scale, it is general
family, mother's education, father's educa- and makes no specific reference to the school
tion, family socioeconomic status, father's experience.
occupational status, mother's occupational To identify the path from global self-es-
status, and numberof siblings. teem to self-estimate of intelligence, the fol-
The model of the reciprocal effects of glo- lowing instrumental variables were used:
bal self-esteem and academic self-esteem fits respondent's height, weight, number of
the data reasonably well, with a chi-square/ friends, level of social support, closeness to
d.f. ratio of 4.05. This shows that global self- father, closeness to mother, parents' ten-
esteem and academic self-esteem affect one dency to reason with the child in disciplin-
another significantly, but that academic self- ary situations, parentalpunitiveness, and sta-
esteem has a more powerful effect on global bility of global self-esteem. All of these
self-esteem than the other way around (.21 would be expected to affect global self-es-
versus .1 1). Both effects are significant at the teem directly and self-estimate of intelli-
p < .05 level. gence only indirectly. Instruments for the
A second question is: Does the same pat- path from perceived intelligence to global
tern of findings appear when we consider a self-esteem include certain performance
specific facet of the self-concept that rests on measures (scores on a job information test
a higher level of abstraction,namely, self-es- and on a test of political knowledge), posi-
timate of intelligence? Facets of the self-con- tive school attitudes, negative school atti-
cept, as Marsh and Shavelson (1985) have tudes, and stability of self-assessment of in-
noted, exist at different levels of specificity. telligence. All of these variables would be
Some facets may be highly specific, others expected to have a direct effect on self-as-

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
150 AMERICAN SOCIOLOGICAL REVIEW

sessment of intelligence and only indirect ef- "meaningful insight ability" (Berger and
fects on global self-esteem. Again, cross- Conner 1969), it is possible that those who
lagged effects were constrained to 0 in order think well of themselves generally may as-
to provide a source of instrumentation. sume they are probably good at insight and
This model was found to fit the data rea- that those who have negative general atti-
sonably well, as shown by a chi-square/d.f. tudes toward themselves will assume that
ratio of 3.45. Again, we find that the specific they probably are not. In this instance, glo-
facet of intelligence has a more powerful ef- bal self-esteem would have a more powerful
fect on global self-esteem than the other way impact on specific self-esteem than the other
around. Both variables produce significant way around.
effects upon one another,but the effect of in-
telligence on global self-esteem is .24,
Is the Degree to Which Specific Academic
whereas the effect of global self-esteem on
Self-Esteem Affects Global Self-Esteem a
intelligence is .11.
Intellectual ability, then, whether narrowly Function of How Much Academic
Performance Is Personally Valued?
conceived as academic ability or broadly
conceived as general intelligence, produces The assertion that the relationship between
a more powerful effect on global self-esteem one's evaluation of a particular facet of the
than the other way around. Note that this ef- self and one's global self-esteem is a func-
fect appears despite the fact that only single tion of the importance of that facet in one's
items are used to measure the two more nar- hierarchy of self-values seems highly plau-
rowly defined concepts (academic self-es- sible. Nevertheless, empirical evidence is
teem and self-estimate of intelligence), while mixed. Our data set allows us to test this as-
global self-esteem is modeled with multiple sertion as it applies to academic self-esteem,
indicators based on a widely used measure. global self-esteem, and valuing academic
High school boys' general feelings of self- performance.
worth appear to be significantly affected by To do this we divided our respondents into
their judgments of their intellectual ability, two groups at the mid-point of the range of
but their general feeling of self-worth has a the variablemeasuringhow much the respon-
weak effect on their opinions of their intel- dent valued academic performance.We then
lectual ability. reran our analyses of the reciprocal effects
We do not suggest, of course, that specific of global and academic self-esteem sepa-
self-esteem is necessarily more likely to af- rately for the respondentsin the low-value (N
fect global self-esteem than the other way = 333) and high-value (N = 1,487) segments
around. Intellectual ability may not be typi- of the distribution.
cal of other facets of the self. Our prediction Our models in both cases were satisfac-
would be that the relative effects of various tory, with a chi-square/d.f. ratio for the low-
facets of the self-concept on global self-es- value group of 1.72 and 2.13 for the high-
teem would depend on the degree of value group. The reciprocal effects between
"schematization" of the specific facets global self-esteem and academic self-esteem
(Markus 1977)-if a particularfacet is firmly were significant only for the high-value
established or solidly crystallized and is bol- group; neither effect was significant in the
stered by clear evidence, then it is likely to smaller, but still substantial, sample that did
be fairly independentof the individual's gen- not particularlyvalue academic performance.
eral feeling of self-worth. For example, if The magnitudeof the effect of academic self-
many experiences of failure have taught a esteem on global self-esteem was more than
pupil through clear evidence that he or she is three times greaterfor those who valued aca-
incompetent in school, then his or her gen- demic performancerelatively highly-the ef-
eral feeling of self-worth will do little to fect for the high-value group was .23 (t =
change this conviction. But if a particular 5.55); the effect for the low-value group was
facet is nonschematic (uncrystallized), then .07 (t =.77). On the other hand, the effect of
global self-esteem may exercise a more pow- global self-esteem on academic self-esteem
erful effect on that specific facet. For ex- was similar for the two groups-the effect
ample, if people are asked to judge their for the high-value group was .12 (t = 2.70);

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 151

for the low-value group the effect was .15 (t tain underprivilegedgroups is that they have
= 1.38). In fact, it is only among the high- low self-esteem and that their performance
value group that specific academic self-es- could be improved by enhancing their self-
teem has a greater effect on global self-es- esteem. But this argumentassumes that poor
teem than the other way around (.23 versus self-esteem causes poor school performance.
.12). Thus, at least in this instance, our find-
If school marks are chiefly responsible for
ings strongly support the conjecture that the self-esteem ratherthan the other way around,
effect of specific self-esteem on global self- then efforts to enhance self-esteem for the
esteem is a function of the degree to which purpose of raising children's achievement
the relevant area of functioning is valued. levels would be misguided.
We also examined whether the level of Second, assuming that self-esteem is caus-
specific self-esteem or global self-esteem af- ally related to school marks, the question re-
fects the degree to which the relevant area of mains: Which self-esteem-specific or glo-
functioning is valued. Using academic self- bal? If, for example, specific self-esteem af-
esteem as our measure of specific self-es- fects school performance but global self-es-
teem, we tested its effect relative to that of teem does not, then the efforts of educators
global self-esteem on valuing academic per- to help children to like and respect them-
formance. In addition, we tested whether selves as a whole may contributeto their psy-
valuing academic performance reciprocally chological well-being but do little to improve
affected academic self-esteem. We did this their school performance.What may actually
by adding "valuing academic performance" be needed is to help them improve their spe-
as a third endogenous variable to our full- cific (academic) self-esteem. It is our impres-
sample model examining the reciprocal ef- sion that most interventions designed to im-
fects of academic and global self-esteem. prove self-esteem and thus to enhance aca-
The new model, which also includes recipro- demic performance are actually attempts to
cal paths between academic self-esteem and modify global rather than specific self-es-
valuing academic performance,seems appro- teem. It is thus essential to know which kind
priately identified and fits the data reason- of self-esteem to change in orderto positively
ably well (chi-square/d.f. = 3.88). Neverthe- affect children's success in school.
less, none of the effects involving academic In an earlier study (Rosenberg, Schooler,
performance was significant, nor were the and Schoenbach 1989), we examined the re-
reciprocal effects between academic and glo- ciprocal effects of global self-esteem and
bal self-esteem notably changed from what school marks. That analysis showed that
they were when valuing academic perfor- school marks had a more powerful effect on
mance was not included in the model. These global self-esteem than global self-esteem
findings seem to rule out the theoretical pos- had on school marks (. 15 versus .08).
sibilities that global and/or academic self-es-Whereas the effect of marks on global self-
teem affect the degree to which academic esteem was highly significant, the effect of
performance is valued or that valuing aca- global self-esteem on marks was only mar-
demic performanceaffects academic self-es- ginally significant. Does this relationship
teem. What valuing academic performance also exist between specific (academic) self-
does is to increase the effect of academic esteem and school marks?
self-esteem on global self-esteem. We constructed a model to test the recip-
rocal effects of academic self-concept and
school performance. The data fit the model
Academic Self-Esteem and Academic reasonably well, yielding a chi-square/d.f.
Performance: Which Causes Which? ratio of 3.5. Each variable, we find, exercises
Low self-esteem has often been invoked as a a strong and statistically significant effect on
possible cause of poor school performance the other; furthermore, the effects are ap-
(e.g., CaliforniaLegislature 1986; Covington proximately equal. The effect of grades on
1984; Purkey 1970; Scheirer and Kraut academic self-esteem is .27; the effect of
1979). For example, it is sometimes sug- academic self-esteem on school marks is .30.
gested that one reason for the relatively poor Similar results occur when we consider the
academic performance of people from cer- reciprocal effects of school marks and self-

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
152 AMERICAN SOCIOLOGICAL REVIEW

assessment of intelligence. These data also nificantly affect self-confidence (beta = .17).
fit the model reasonably well, with a chi- The effect of self-confidence on school
square/d.f. ratio of 3.25. Each variable has a marksis lower (.07) and not significant. Nei-
highly significant effect on the other and ther the effect of grades on self-deprecation
again these effects are equal (both effects are (-.11) nor the effect of self-deprecation on
.36); self-estimates of intelligence thus ap- grades (-.04) was significant, although the
pear to affect, and to be affected by, school former was at the borderline of significance
performance. This evidence suggests that (p < .08). Not surprisingly, given Owens's
changes in self-esteem can in fact bring (1993) finding that self-confidence is more
about an improvement in school perfor- strongly related to global self-esteem than is
mance; but it must be the right self-esteem, self-deprecation, the magnitudes of the paths
namely, specific academic self-esteem. Strat- between self-confidence and grades are also
egies designed to enhance global self-esteem more similar to the magnitudes of the paral-
(for example, by convincing a child that he lel paths between global self-esteem and
or she is lovable or worthy of respect) will, school marks we reportedin our earlier study
our data suggest, have little or no effect on (Rosenberg, Schooler, and Schoenbach
his or her school performance. 1989). In any case, all of these effects are
It could also be argued that, to improve smaller than the parallel effects involving
school performance, positive or negative academic self-esteem or self-assessment of
self-esteem should be changed rather than intelligence. Furthermore, although our
global self-esteem. On the one hand, efficacy analyses suggest that raising the academic
theory implies that it is primarily the self- self-esteem or self-estimates of intelligence
confidence generated by positive self-esteem among students may increase academic per-
that enhances school performance (Bandura formance, nothing in our findings suggests
1982). On the other hand, the literature on that raising their levels of global self-esteem
learned helplessness suggests that the self- or general self-confidence, or decreasing
deprecation of negative self-esteem, feelings their levels of general self-deprecation would
of personal helplessness, and poor perfor- have any such effect.
mance are causally linked (Abramson,
Seligman, and Teasdale 1978).
SUMMARY AND DISCUSSION
To test whether either self-confidence or
self-deprecation is more closely tied to aca- Our aim has been to better understand the
demic performance than is global self-es- nature, interconnections, and relevance of
teem, we replicated our analyses substituting specific self-esteem and global self-esteem.
each of these variables for global self-es- In evaluating our conclusions we must ad-
teem.5 The only significant finding that dress several limitations in our data. First,
emerged regarding the relationship between our only measure of global self-esteem is the
these two aspects of self-esteem and aca- Rosenberg Self-Esteem Scale (1989). Al-
demic performance is that school marks sig- though it is the most widely used of all self-
esteem measures, it is only one among many.
I Some of the other available measures may
Self-confidence and self-deprecation are
based on structural equation measurement mod- include somewhat different aspects of self-
els derived from the original Kohn and Schooler esteem than does the Rosenberg scale.6
(1983) analyses of self-esteem. Four indicators of
6 For example, the Coopersmith Self-Esteem
self-confidence are based on responses to the fol-
lowing four statements (coded from 1 through 5, Inventory (Coopersmith 1967) contains four in-
never true to almost always true): "I am a person ternally consistent subscales (social self, peers,
of worth"; "I feel that I have a number of good home/parents, academic), although factor analy-
qualities"; "I am able to do things as well as most sis revealed a dominant global self-esteem factor
other people"; "I take a positive attitude toward (Burns 1979; Edgar, Powell, Watkins, Moore, and
myself." The two indicators of self-deprecation Zakharov 1974). The Tennessee Self-Concept
are based on responses to two statements (same Scale (Fitts 1965) includes, among others, sub-
coding as above): "Sometimes I think I am no scales on how respondents view their own moral
good at all"; "I feel that my life is not very use- worth and perceive themselves in reference to
ful." their families. Analyses of the Piers-Harris Scale

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 153

Another limitation of our data is that our more cognitive component and tends to be
only specific self-esteem measure is aca- more strongly associated with behavior or
demic self-esteem. We do not know if the behavioral outcomes.
same patternof results would hold if we used (3) We have suggested that global and spe-
measures of other types of specific self-es- cific self-esteem may each mediate the effect
teem, like social or sexual self-esteem. Dif- of the other. It was not possible to test
ferent kinds of specific self-esteem vary whether the relationship between specific
across a wide range of potentially relevant self-esteem and psychological well-being
dimensions. One importantdimension is the was mediated by global self-esteem because
degree of specificity. Thus, although aca- the original relationshipwas so weak, but our
demic self-esteem is more specific than over- data are consistent with the view that the ef-
all self-esteem, it is less specific than math- fect of global self-esteem on behavioral out-
ematical self-esteem. As we have seen, a come is mediated by its effect on academic
general tenet of many research findings on self-esteem.
attitudes is that the more specific the attitude, (4) We have considered the reciprocal ef-
the more powerful a predictor it is of the re- fects of global and specific self-esteem on
lated specific behavior (e.g., Fishbein and one another.We pointed out that specific fac-
Azjen 1975). Given that self-esteem is an at- ets of the self may vary in level of abstrac-
titude, the more specific the self-esteem is, tion. Our empirical findings suggest that
the more accurately it should predict relevant more specific forms of self-esteem, such as
behavior. As our own findings indicate, spe- academic self-esteem or self-assessment of
cific academic self-esteem has a greater ef- intelligence, tend to have greater effects on
fect on school marks than does global self- more global forms of self-esteem than more
esteem. global forms of self-esteem have on more
Despite these limitations, we believe that specific ones.
our findings are important: (5) We have found that, at least among the
(1) We have documented the obvious but variables we investigated, the effect of spe-
frequently overlooked fact that global and cific self-esteem on global self-esteem is af-
specific self-esteem are neither equivalent fected by the degree to which the relevant
nor interchangeable, and that one cannot au- role or behavior is personally valued. Our
tomatically serve as a surrogatefor the other. analyses demonstratedthat the effect of spe-
(2) Focusing on academic self-esteem as cific academic self-esteem on global self-es-
an example of specific self-esteem, we have teem is a function of how highly academic
demonstrated that global and specific self- performance is valued. Furthermore, the
esteem have decidedly different correlates. value placed on academic performanceis not
Global self-esteem appears to be heavily af- a function of either academic or global self-
fective in nature and tends to be associated esteem.
with overall psychological well-being. Spe- (6) We also considered the reciprocal ef-
cific self-esteem, in contrast, being more fects of self-esteem and school marks, an im-
judgmental and evaluative, appears to have a portant behavioral outcome among school
populations. The data show that the nature
(Piers and Harris 1964) have consistently re- and direction of these effects depends on
vealed six interpretable factors, including popu- whether global or specific self-esteem is con-
larity and social behavior, although no general sidered. School marks, we find, do produce
factor has emerged (Burns 1979). Regardless of an effect on self-esteem, whether we con-
whether each of these scales contains specific sider academic self-esteem, global self-es-
components not covered by the Rosenberg scale, teem, or self-confidence. However, global
the Rosenberg scale would seem to be a more di- self-esteem has very little effect on marks,
rect measure of global self-esteem; the self-evalu- whereas specific self-esteem (academic self-
ation of various aspects of functioning covered by
the Coopersmith, Fitts, and Piers-Harris scales
esteem) has a strong effect on school perfor-
being conceptually akin to more specific forms of mance. Self-esteem, it appears, does affect
self-esteem, such as educational self-esteem, school performance, but it must be the right
whose relationship to global self-esteem is a fo- kind of self-esteem, namely, specific self-es-
cus of the present paper. teem. This last finding is particularly ironic,

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
154 AMERICAN SOCIOLOGICAL REVIEW

as educators and policymakers almost invari- REFERENCES


ably focus on precisely the wrong type of
self-esteem-global self-esteem-when they Abramson, Lyn, Martin Seligman, and John D.
introduce interventions to improve students' Teasdale. 1978. "Learned Helplessness in Hu-
mans: Critique and Reformulation."Journal of
performance in school. Understanding the
Abnormal Psychology. 87:49-74.
nature and interrelationships of global and Bachman, Jerald G. 1970. Youth in Transition.
specific self-esteem would, thus, seem more Vol. 2. The Impact of Family Background and
than a mere academic exercise. Heeding the Intelligence on Tenth-GradeBoys. Ann Arbor,
distinction might actually increase the effec- MI: Institute for Social Research.
tiveness of future educational interventions. . 1975. Youth in Transition. Documenta-
tion Manual. Vol. 1. Ann Arbor, MI: Institute
Morris Rosenbergwas the central sociological for Social Research.
researcherand theoriston self-esteemand con- Bandura, Albert. 1982. "The Self and Mecha-
ceptionsof the self andprobablyone of the lead- nisms of Agency." Pp. 3-39 in Psychological
ing expertson these issues in any discipline.He Perspectives on the Self, vol. 1, edited by J.
was a Professorof Sociologyat the Universityof Suls. Hillsdale, NJ: Erlbaum.
Marylandand at varioustimeswas a memberof Baumeister, Roy F. 1982. "A Self-Presentational
theLaboratoryof Socio-Environmental Studiesof View of Social Phenomena." Psychological
the NationalInstituteof MentalHealth.He was Bulletin 91:3-26.
Vice Presidentof the AmericanSociologicalAs- Beck, Aaron T., A. John Rush, Brian F. Shaw,
sociationand the winnerof manyawardsinclud- and G. Emery. 1979. Cognitive Therapy of De-
ing the AAASSociopsychologicalPrize, the ASA pression. New York: Guilford.
Social PsychologySection CooleyMeadAward, Berger, Joseph and Thomas L. Conner. 1969.
and theASADistinguishedContribution to Schol- "Performance Expectations and Behavior in
arshipAward.MannyRosenbergdied on Febru- Small Groups."Acta Sociologica 12:186-98.
ary 14, 1992. Burns, Robert B. 1979. The Self Concept. New
Carmi Schooler is Acting Chief of the Labora- York: Longman Group Limited.
tory of Socio-Environmental Studiesof the Na- California Legislature. 1986. Assembly Bill No.
tional Instituteof MentalHealth.His sociologi- 3659, chap. 1065. A Bill to Establish Califor-
cal researchfocuses on the social structuraland nia Commission to Promote Self-Esteem and
culturaldeterminantsof psychologicalfunction- Personal and Social Responsibility.
ing throughthe life span.He is presentlyinvolved Coopersmith, Stanley. 1967. The Antecedents of
in a follow-upstudyof respondentsoriginallyin- Self-Esteem. San Francisco, CA: Freeman.
terviewedin 1964 and 1974for a studyhe con- Covington, Martin. 1984. "The Motive for Self-
ducted with Melvin Kohnon the Psychological Worth." In Research on Motivation in Educa-
effectsof occupationalconditions. tion, vol. 1: Student Motivation, edited by R.
Ames and C. Ames. New York: Academic
Carrie Schoenbach is former ResearchAssoci- Press.
ate, now guest worker, in the Laboratoryof Crocker, Jennifer and Brenda Major. 1989. "So-
Socio-Environmental Studiesof the NationalIn- cial Stigma: The Affective Consequences of
stituteof MentalHealth.She has beena collabo- Attributional Ambiguity." Journal of Person-
rator with CarmiSchoolerand MelvinKohnin ality and Social Psychology 60:218-28.
analyzingthe effectsof social-structuraland cul- Cross, William E. 1985. "Black Identity: Redis-
turalconditions-particularlytheeffectsof occu- covering the Distinction Between Personal
pational conditions,social stratificationand so- Identity and Reference Group Orientation." In
cial class positions-on valuesandpsychological Beginnings: The Social and Affective Develop-
functioning. ment of Black Children, edited by M. B. Spen-
FlorenceRosenberghas beena SeniorResearch cer, G. K. Brookins, and W. R. Allen.
Associatein the Departmentof MilitaryPsychia- Hillsdale, NJ: Erlbaum.
try, WalterReedArmyInstituteof Researchand Diggory, James C. 1966. Self-Evaluation: Con-
the Departmentof Psychiatry, UniformedSer- cepts and Studies. New York: Wiley.
vices Universityof theHealthSciences.Hermain Edgar, Patricia, R. J. Powell, D. Watkins, R. J.
researchinteresthas beenon stress, coping,and Moore, and Olive Zakharov. 1974. An Analy-
mental health amongArmywives and families, sis of the Coopersmith Self-Esteem Inventory."
focusing on the perceptualfactors that contrib- Australian Psychology 9:52-63.
ute to satisfactionwithlife.Atpresent,as a Guest Faunce, William A. 1982. "The Relation of Sta-
Scientistat the WalterReedArmyInstitute,she is tus to Self-Esteem: Chain Saw Sociology at the
continuingresearch on this subject with other Cutting Edge." Sociological Focus 15:163-78.
populations. Fishbein, Martin and Icek Ajzen. 1975. Belief,

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
GLOBAL SELF-ESTEEM AND SPECIFIC SELF-ESTEEM 155

Attitude, Intention and Behavior. Reading, into the Impact of Social Stratification.
MA: Addison-Wesley. Norwood, NJ: Ablex.
Fitts, William H. 1965. Tennessee Self-Concept Luck, Patrick and Jerold Heiss. 1972. "Social De-
Scale. Nashville, TN: Counselor Recordings terminantsof Self-Esteem in Adult Males." So-
and Tests. ciology and Social Research 57:69-84.
Gecas, Viktor and Michael Schwalbe. 1983. "Be- Major, Brenda, Anne Marie Sciacchitano, and
yond the Looking-Glass Self: Social Structure Jennifer Crocker. 1993. "In-Group vs. Out-
and Efficacy Based Self-Esteem." Social Psy- Group Comparisons and Self-Esteem." Person-
chology Quarterly 46:77-88. ality and Social Psychology Bulletin 19:711-
Gecas, Viktor and Monica A. Seff. 1990. "Social 21.
Class and Self-Esteem: Psychological Central- Markus, Hazel. 1977. "Self-Schemata and Pro-
ity, Compensation, and the Relative Effects of cessing Informationabout the Self." Journal of
Work and Home." Social Psychology Quar- Personality and Social Psychology 35:63-78.
terly 53:165-73. Marsh, Herbert W. 1986. "Global Self-Esteem:
Greenwald, Anthony G. 1980. "The Totalitarian Its Relation to Specific Facets of Self-Concept
Ego: Fabricationand Revision of Personal His- and Their Importance."Journal of Personality
tory." American Psychologist 35:603-18. and Social Psychology 51:1224-236.
Harter, Susan. 1985. "Competence as a Dimen- . 1990. "Influences of Internal and Exter-
sion of Self-Evaluation: Toward a Comprehen- nal Frames of Reference on the Formation of
sive Model of Self-Worth." Pp. 55-121 in The Math and English Self-Concepts." Journal of
Development of the Self, edited by R. L Leahy. Educational Psychology 82:107-16.
Orlando, FL: Academic Press. Marsh, Herbert W. and Richard J. Shavelson.
Hoelter, Jon. 1986. "The Relationship Between 1985. "Self-Concept: Its Multifaceted, Hierar-
Specific and Global Evaluations of the Self: A chical Structure." Educational Psychologist
Comparison of Several Models." Social Psy- 20:107-25.
chology Quarterly 49:129-41. Maslow, Abraham H. 1970. Motivation and Per-
Hoge, Dean R. and John D. McCarthy. 1984. "In- sonality. Rev. ed. New York: Harper.
fluence of Individual and Group Identity Sa- McCarthy,John D. and William L. Yancey. 1971.
lience in the Global Self-Esteem of Youth." "Uncle Tom and Mr. Charlie: Metaphysical
Journal of Personality and Social Psychology Pathos in the Study of Racism and Personal
47:403-14. Disorganization."American Journal of Sociol-
James, William. [1890] 1950. The Principles of ogy 76:648-72.
Psychology. New York: Dover. Menaghan, Elizabeth and Toby L. Parcel. 1990.
Jones, Stephen C. 1973. "Self and Interpersonal "Determining Children's Home Environments:
Evaluations: Esteem Theories Versus Consis- The Impact of Maternal Characteristics and
tency Theories." Psychological Bulletin 79: CurrentOccupational and Family Conditions."
185-99. Department of Sociology, Ohio State Univer-
Joreskog, Karl G. and Dag Sorbom. 1976. "Sta- sity, Columbus. OH. Unpublished manuscript.
tistical Models and Methods for Analysis of Owens, Timothy J. 1993. "Accentuate the Posi-
Longitudinal Data." Pp. 285-325 in Latent tive-and the Negative: Rethinking the Use of
Variables in Socioeconomic Models, edited by Self-Esteem, Self-Deprecation and Self-Confi-
D. J. Aigner and A. S. Goldberger. Amsterdam, dence." Social Psychology Quarterly 56:288-
The Netherlands: North Holland. 99.
Kaplan, Howard B. 1975. Self-Attitudes and De- Pearlin, Leonard I. and Morton A. Lieberman.
viant Behavior. Pacific Palisades, CA: Good- 1979. "Social Sources of Emotional Distress."
year. Pp. 217-48 in Research in Community and
Kaplan, Howard B. and Alex D. Pokorny. 1969. Mental Health, vol. 1, edited by R. G.
"Self-Derogation and Psychosocial Adjust- Simmons. Greenwich, CT: JAI.
ment." Journal of Nervous and Mental Disease Pearlin, Leonard I. and Carmi Schooler. 1978.
149:421-34. "The Structure of Coping." Journal of Health
Kitano, Harry H. 1989. "Alcohol and Drug Use and Behavior 19:2-21.
and Self-Esteem: A Sociocultural Perspective." Piers, Ellen V. and Dale B. Harris. 1964. "Age
Pp. 294-326 in The Social Importance of Self- and Other Correlates of Self-Concept in Chil-
Esteem, edited by A. Mecca, N. Smelser, and dren. Journal of Educational Psychology
J. Vasconcellos. Berkeley, CA: University of 55:91-95.
California Press. Porter, Judith R. and Robert E. Washington.
Kohn, Melvin L. and Carmi Schooler. 1969. 1989. "Developments in Research on Black
"Class, Occupation and Orientation."American Identity and Self-Esteem: 1979-1988." Revue
Sociological Review 34:659-78. Internationale de Psychologie Sociale T2:339-
. 1983. Workand Personality: An Inquiry 53.

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions
156 AMERICAN SOCIOLOGICAL REVIEW

Purkey, William W. 1970. Self-Concept and Scheirer, Mary A. and Robert E. Kraut. 1979. "In-
School Achievement. Englewood Cliffs, NJ: creasing Educational Achievement Via Self-
Prentice-Hall. Concept Change." Review of Educational Re-
Rosenberg, Morris. 1979. Conceiving the Self. search 49:131-49.
Malabar, FL: Robert E. Krieger. Shrauger, Sidney J. 1972. "Self-Esteem and Re-
. 1985. "Self-Concept and Psychological actions to Being Observed by Others."Journal
Well-Being in Adolescence." Pp. 205-46 in of Personality and Social Psychology 23:193-
The Development of the Self, edited by R. L. 200.
Leahy. Orlando, FL: Academic Press. Swann, William. 1987. "Identity Negotiation:
. 1989. Society and the Adolescent Self-Im- Where Two Roads Meet." Journal of Person-
age. Rev. ed. Middletown, CT: Wesleyan Uni- ality and Social Psychology 53:1038-51.
versity Press. Swann, William, Alan Stein-Seroussi and R.
Rosenberg, Morris and Leonard I. Pearlin. 1978. Brian Giesler. 1992. "Why People Self-
"Social Class and Self-Esteem among Children Verify." Journal of Personality and Social Psy-
and Adults." American Journal of Sociology chology. 62:392-401.
84:53-77. Tesser, Abraham and Jennifer Campbell. 1983.
Rosenberg, Morris, Carmi Schooler, and Carrie "Self-Definition and Self-Evaluation Mainte-
Schoenbach. 1989. "Self-Esteem and Adoles- nance." Pp. 1-31 in Psychological Perspectives
cent Problems: Modeling Reciprocal Effects." on the Self, vol. 2, edited by J. Suls and A. G.
American Sociological Review 54:1004-18. Greenwald. Hillsdale, NJ: Erlbaum.
Rosenberg, Morris and Roberta G. Simmons. Thomas, Michael. 1989. Zentralitat und Selbst-
1972. Black and White Self-Esteem. Washing- konzept (Centrality and Self-Concept). Bern,
ton, DC: American Sociological Association. Switzerland: Hans Huber.
Scheff, Thomas, Suzanne M. Retzinger, and Weick, Karl E. 1964. "Reduction of Cognitive
Michael T. Ryan. 1989. "Crime, Violence, and Dissonance throughTask Enhancementand Ef-
Self-Esteem: Review and Proposals." Pp. 165- fort Expenditure." Journal of Abnormal and
99 in The Social Importance of Self-Esteem, Social Psychology 68:533-39.
edited by A. M. Mecca, N. J. Smelser, and J. Wylie, Ruth C. 1979. The Self-Concept. Vol. 2:
Vasconcellos. Berkeley, CA: University of Theory and Research on Selected Topics. Rev.
California Press. ed. Lincoln, NE: University of Nebraska Press.

This content downloaded from 61.129.42.30 on Fri, 08 May 2015 08:44:25 UTC
All use subject to JSTOR Terms and Conditions

You might also like