Citation
A study of self-esteem, academic self-concept and academic achievement of African American students in grades five, seven, and ten in a predominately white suburban school district

Material Information

Title:
A study of self-esteem, academic self-concept and academic achievement of African American students in grades five, seven, and ten in a predominately white suburban school district
Creator:
Thurman, Mary Fleming
Publication Date:
Language:
English
Physical Description:
113 leaves : ; 28 cm

Subjects

Subjects / Keywords:
African American students -- Psychology ( lcsh )
Academic achievement -- United States ( lcsh )
Self-esteem ( lcsh )
Self-perception ( lcsh )
Suburban schools -- Social conditions -- United States ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 107-113).
General Note:
School of Education and Human Development
Statement of Responsibility:
by Mary Fleming Thurman.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
45139807 ( OCLC )
ocm45139807
Classification:
LD1190.E3 2000d T48 ( lcc )

Full Text
A STUDY OF SELF-ESTEEM, ACADEMIC SELF-CONCEPT AND
ACADEMIC ACHIEVEMENT OF AFRICAN AMERICAN STUDENTS
IN GRADES FIVE, SEVEN, AND TEN IN A PREDOMINANTLY
WHITE SUBURBAN SCHOOL DISTRICT
by
Mary Fleming Thurman
B.S., Tuskegee University, 1971
M.S., Troy State University, 1988
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Educational Leadership and Innovation
2000


2000 by Mary Fleming Thurman
All rights reserved.


This thesis for the Doctor of Philosophy
degree by
Mary Fleming Thurman
has been approved
by
Rod Muth

2 O/j c>
Date


DEDICATION
I dedicate this thesis to the memories of Lela, my
mother, and Major, my father. Their lessons about life
sustained me many times as I completed my dissertation.
I love you mama and dad.


ACKNOWLEDGEMENTS
A significant portion of the past few years of my life
has been spent in sequestration as I completed my
dissertation. I offer my sincere thank you to those
who made significant contributions to my research
project and contributed to its success:
Professor Alan Davis, my advisor, for his wealth of
suggestions and ideas, time, and commitment to my work;
Professors Rod Muth, Nadyne Guzman, A1 Ramirez, and
Margaret A. Bacon, my graduate committee, for their
helpful suggestions and constructive review of my
thesis;
Karen Bertel for her invaluable assistance in document
research;
Principals, assistant principals, counselors, and staff
in the suburban school district where the study took
place for their assistance in recruiting students for
this study;
Students and parents in this suburban school district
for their assistance;
Daniel, my husband, for his love, his steadfast support
of my needs, and his constant reassurance and
encouragement;
Sherry and Danielle, my daughters, and Sylvia, my
mother-in-law for their love and encouraging words.


Thurman, Mary Fleming (Ph.D., Educational Leadership
and Innovation)
A Study of Self-Esteem, Academic Self-Concept and
Academic Achievement of African American Students
in Grades Five, Seven, and Ten in a Predominantly
White Suburban School District
Thesis directed by Professor Alan Davis
ABSTRACT
This study examined the relationship between
self-esteem, academic self-concept, and academic
achievement among African American students. Data were
gathered from African American students who were in
grades five, seven, and ten in a predominantly white,
middle class, suburban school district. Both
quantitative and qualitative research approaches were
used to collect data using Marsh's Self Description
Questionnaire (SDQ), report card grade averages, Iowa
Test of Basic Skills (ITBS) scores, and student
interviews.
Findings revealed that although African American
students in this study reported high self-esteem at
each grade level self-esteem did not increase across
grade levels as hypothesized. Self-esteem scores of
students in grade 5 were higher than middle and high
school students. The hypothesis that elementary
students would have a stronger relationship between
academic achievement and self-esteem scores was not
Vi


supported. No correlation existed between academic
achievement and self-esteem at any grade level. The
hypothesis that high school students would have a
stronger relationship between their academic
self-concept scores ahd achievement than middle and
elementary students was partially supported. Students
in grade 10 showed a significant correlation between
academic self-concepts scores and achievement as
measured by semester report card grade averages whereas
no significant correlation between academic self-
concept and achievement was obtained at grade 7 or
grade 5.
Results from student interviews revealed that the
determinants of self-esteem, in relation to African
American students, are not associated with academic
achievement, but rather relate to peer acceptance,
participation in sports or other nonacademic factors.
An important finding from student interviews revealed
that academic self-concept of ability may produce a
more positive correlation between academic achievement
than global self-esteem.
This abstract accurately represents the content of the
candidate's thesis. I recommend its publication.
Signed
Alan Davis
vii


CONTENTS
Tables ...................................... x
CHAPTER
1. NATURE OF THE STUDY......................... 1
General Problem......................... 1
Background of the Problem............... 5
Theoretical Framework.................. 10
Problem Focus.......................... 14
Limitations of the Study............... 20
Definition of Terms.................... 20
Summary................................ 21
2. REVIEW OF THE LITERATURE................... 22
Self-esteem/self-concept............... 23
Related Research....................... 30
Self-esteem and
Academic Achievement................... 33
Academic Self-concept and
Academic Achievement................... 42
viii


Self-esteem and African American
Student Achievement ................... 46
3. METHODOLOGY............................... 55
Sample................................. 57
Data Collection........................ 58
The Self Description
Questionnaire ..................... 59
Iowa Test of Basic Skills.......... 62
Students' Grade Report ............ 65
Student Interviews ................ 66
Statistical Analysis............... 68
4. ANALYSIS OF FINDINGS....................... 72
Summary............................ 83
The Interviews ........................ 84
5. CONCLUSIONS AND RECOMMENDATIONS............ 93
APPENDIX
A. SELF DESCRIPTION QUESTIONNAIRE............ 102
B. STUDENT INTERVIEW QUESTIONS .............. 105
REFERENCES.......................................... 107
ix


TABLES
Table
4.1 Study participants by grade .................. 75
4.2 Self-esteem scores as
measured by the SDQ........................... 76
4.3 Differences in self-esteem
expressed in effect sizes .................... 78
4.4 Correlations of self-esteem
with academic achievement..................... 79
4.5 Correlations of academic
self-concept with
academic achievement.......................... 81
4.6 Reading self-concept scores
correlated with actual reading
grades and ITBS reading scores................ 82
4.7 Mathematics self-concept scores
correlated with actual
mathematics grades and
ITBS mathematics scores ...................... 83
x


CHAPTER 1
NATURE OF THE STUDY
The purpose of this study is to examine the
relationship between self-esteem, academic self-concept,
and academic achievement among African American students.
The specific question is: What is the relationship
between self-esteem, academic self-concept, and academic
achievement among African American students in a
predominantly white suburban school district?
The General Problem
Extensive research has suggested a positive
relationship between academic achievement and self-esteem
(Gwin, 1990; Justice, Lindsey, & Morrow, 1999; Maruyama,
Rubin, & Kingsbury, 1981; Mayo-Booker & Gibbs, 1997;
Purkey, 1970; Simmons, Brown, Bush, & Blyth, 1978;
Voelkl, 1993). According to Voelkl (1993), however, this
relationship is not consistent in studies of African
American students. Despite lower levels of achievement
as compared to whites, African Americans expressed higher
average levels of self-esteem (Voelkl, 1993).
1


Minority students' standardized achievement test
scores at the state and national levels are lower than
their white counterparts (NAEP, 1992), however; their
level of self-esteem is higher (Madhere, 1991; Simmons
et al., 1978). In the district used for this study,
minority students consistently score below the district,
state, and national averages on standardized tests.
Other researchers (Austin & Garber, 1982; Ferguson, 1990;
Fischer et al., 1996; Fordham & Ogbu, 1986; Hilliard,
1991; Irvine, 1990; Lee, Winfield, & Wilson, 1991;
Miller, 1995; Ogbu, 1978; Voelkl, 1993; Walton & Taylor,
1996/1997) indicate that the minority-white achievement
gap is consistent with standardized achievement test
scores and in all areas of academic achievement.
Austin and Garber (1982) discuss the rise and fall
of test scores in general. They contend that African
American students have lower academic achievement because
they are frequently subjected to racism and a
disproportionate number live in poverty. Ogbu (1978)
explored the caste system to explain the gap in African
American students' performance compared to their white
counterparts. Ogbu contends that the cognitive,
linguistic, and motivational attributes of African
American students may be adversely affected due to the
2


job ceiling and other caste barriers that are imposed.
African American students' low academic achievement might
result from these students assuming that job
opportunities and advancements will not be available to
them, while parents encourage their children to do well
in school, Ogbu's research revealed that African American
children saw that success in school did not equate to
better jobs or better living conditions.
Fischer et al. (1996) contend that the caste system
greatly contributes to low test scores and achievement
among minorities. In a model of how low ethnic position
causes low test scores, Fischer et al. argue that "a
racial or ethnic group's position in society determines
its measured intelligence rather than vice versa"
(p. 173). They indicate that subordination leads to low
test scores and low academic performance through
socioeconomic deprivation, group segregation, and the
stigma of inferiority. Despite these findings (low test
scores and low academic achievement among minority
students), African American students' self-esteem appears
to be higher than their white counterparts.
Coopersmith (1979) states that self-esteem is "the
evaluation which the individual makes and customarily
maintains with regard to himself: it expresses an
3


attitude of approval or disapproval, and indicates the
extent to which the individual believes himself to be
capable, significant, successful, and worthy" (p. 5).
Coopersmith gives a global definition of self-esteem.
Numerous researchers studying the relationship between
self-esteem and academic achievement give a global or
general definition of self-esteem that is similar to
Coopersmith's definition (Gergen 1971; Maruyama et al.,
1981; Purkey, 1970; Rosenberg 1965; Simmons et al.,
1978). They describe self-esteem in terms of self-worth.
Other researchers suggest that a more positive
correlation would be found by studying the relationship
between self-esteem and academic achievement using a more
specific measure of self-esteem. One researcher with
this belief is Herbert Marsh.
Marsh (1986) contends that self-esteem can be more
adequately understood when its multidimensionality is
included. Marsh contends that if a specific content area
measure is not included students will report a higher
self-esteem. Marsh (1992) further contends that in order
to determine more adequately the relationship between
self-esteem and academic achievement a measure specific
to academic subjects must be used.
4


The Background of the Problem
A number of researchers indicate that self-esteem
should relate to academic achievement. Research
conducted by Harter (1982) and Lay and Wakstein (1985)
contends that higher achieving students will report
higher self-esteem than lower achieving students.
Research studies investigating students' self-esteem and
academic achievement show a positive correlation
(Thomas-Bentley, 1988; Wylie, 1979). In other words,
students who perceive that they can achieve at high
levels and do well in school actually do achieve at high
levels.
Osborne (1995) found that, according to the
symbolic interactionist perspective "positive feedback in
school would lead to more positive self-evaluations,
whereas negative feedback would lead to more negative
self-evaluations" (p. 449). The symbolic interactionist
perspective refers to the "looking glass" self. It
refers to how a person believes others view her or him.
Therefore, as Osborne stated a student receiving positive
feedback in school is more likely to have positive
feelings of self. Whereas, a person receiving negative
feedback in school will have negative feelings of self in
5


relation to those subjects where the negative feedback is
given.
Lay and Wakstein (1985) posited that students who
do poorly in school will report a lower self-esteem when
compared to someone who is achieving at a higher level.
Their research predicted a positive correlation with
academic achievement and self-esteem. They suggested
that "students who achieve less tend to have lower
self-esteem" (p. 53).
According to Voelkl (1993), research studies
indicate that this correlation of self-esteem and
academic achievement holds true for white students.
However, other research studies report that a paradox
exists for African American students. African American
students have high self-esteem but their academic
achievement is low (p. 42).
Reviewing over 140 studies related to self-esteem
and achievement, Graham (1994) indicates that, in spite
of low academic achievement African American students
have high self-esteem and high expectations for the
future. These studies used survey instruments as one
means to collect data to explain the self-esteem and
achievement paradox of African American students.
6


Research studies involving self-esteem cure usually
done by making comparisons between African American
students' responses and white students' responses to a
survey questionnaire. Bachman and O'Malley (1984) used
several survey instruments, including the Rosenberg
self-esteem scale, to analyze black-white self-esteem
differences. They believe that African Americans' higher
self-esteem in relation to their white counterparts is
due to African Americans using extreme response
categories on questionnaire items. The questionnaire
items all used Likert-type response scales.
Based on their research findings, African Americans
consistently reported the "positive 'agree' end of agree-
disagree scales ..." (Bachman & O'Malley, 1984, p.
626). These researchers argue that the Likert-type
response may be invalid for African American students
because these students cure more likely to use the extreme
responses. They argue that, when self-esteem scores use
a truncated scoring method to control for extreme
responses, the black-white discrepancy disappears
(p. 624). Bachman and O'Malley analyzed extreme response
style using five nationwide samples that employed
responses to a questionnaire.
7


Bachman and O'Malley analyzed a sixth sample that
used face-to-face interviews and found no clear
replication of the findings from the questionnaire
responses. They conclude that, even though no important
response style differences occurred when using
interviews, self-esteem issues should be explored further
using this method to collect data. "In particular,
future analysts may wish to explore whether and how the
race of the interviewer interacts with the race of the
respondent in their relationships with response style"
(p. 635).
According to Crocker and Major (1989), the apparent
discrepancy in self-esteem (i.e., African American
students with higher levels of self-esteem than white
students) can be explained by the different reference
groups used by white and African American students in
making self-comparisons. Crocker and Major contend that
African Americans may compare their outcomes (i.e.,
academic achievement) with those of other African
Americans who are considered the ingroup, rather than to
white students whom they consider the "advantaged
outgroup" (p. 608).
8


Research studying the relationship between
self-esteem and academic achievement typically focuses on
general components of self-esteem. In Marsh's (1992)
research, an academic self-concept measure was used to
determine the relationship between self-esteem and
academic achievement. Marsh contends that higher
correlations will exist between self-concept and academic
achievement when measures specific to academic
self-concept in specific content areas are used.
using Shavelson and Bolus' (1982) multifaceted,
hierarchical model, Marsh, Byrne, and Shavelson (1988)
investigated self-esteem in relation to academic
achievement in specific content areas. Results indicated
that verbal and mathematics achievement were highly
correlated with academic self-concepts in these specific
content areas. Mathematics achievement also had a strong,
positive correlation with mathematics self-concept. The
same results were found for verbal achievement and verbal
academic self-concept. Results further revealed that
mathematics achievement had a weak or nonexistent
correlation with verbal self-concept.
Comparing academic achievement in a specific
subject to academic self-concept in the same subject
produced strong, positive relations. Marsh's research
9


led to the development of the Marsh's Self Description
Questionnaire (SDQ) which was designed to measure a
student's self-concept in a multifaceted manner to
include reading, mathematics, general school, physical
abilities, appearance, peer relations, and parent
relations. Marsh (1990b) indicates that the general self
scale was later included (p. 1). Marsh has since
developed subsequent SDQ instruments to include
additional dimensions that are specific to certain age
groups. This research study uses Marsh's SDQ to study
the relationship between self-esteem, academic
self-concept, and academic achievement.
Studying the relationship between self-esteem,
academic self-concept, and academic achievement using
Marsh's SDQ will provide more insight about the paradox
that exists for African American students. African
American students have high self-esteem, yet their
academic achievement is low (Voelkl, 1993, p. 42).
Theoretical Framework
Osborne (1995) examined the academic achievement
and self-esteem of African American students using
Steele's (1992) theory of disindentification as his
conceptual framework. Steele's theory of
disidentification was used by Osborne to account for both
10


African American students' poor academic performance and
their paradoxically high self-esteem. Osborne's (1995)
research tested three predictions that form the basis for
Steele's disindentification theory:
(a) that African American students score lower on
measures of academic achievement than White
students, (b) that African American students tend
not to report lower self-esteem than White
students, and (c) that correlations between
measures of academic achievement and global
self-esteem should be moderate and significant for
both White and African Americans early in their
education, but that over time the correlations
should weaken for African Americans as they
disidentify. (p. 451)
Steele (1997) makes an analogy of two students in a
classroom: one African American student and one white
student. They are in a better classroom. They have the
same teacher. Their treatment in class is the same. He
queries: "Is it possible, then, that they could still
experience the classroom differently, so differently in
fact as to significantly affect their performance and
achievement there" (p. 613)? Steele contends African
American students' susceptibility to stereotypes (e.g.,
blacks perform poorer than whites in all areas of
academic achievement) causes them to dissociate their
academic achievement from their overall self-esteem.
11


Steele states,
There is the persistent finding that although Black
students underperform in relation to White students
on school outcomes from grades to standardized
tests, their global self-esteem is as high or
higher than that of White students. For both of
these facts to be true, some portion of Black
students must have acquired an imperviousness to
poor school performance. (p. 623)
This imperviousness, Steel contends, is the source of
Black students' disidentification.
Hare and Costenell (1985) measured students' global
self-esteem and their academic achievement in relation to
their home life, school, and peer-group relations. They
found that African American students' academic
performance was below that of white students, but they
had comparable levels of global self-esteem. These
researchers found that, even though African American
students' global self-esteem was comparable their home
and school self-esteem were lower. African American
students' peer-group self-esteem was higher than white
students'. Steele (1997) contends that the results can
perhaps be explained by the disidentification theory.
Steele further states that African American students'
ability stigmatization (e.g., poor performance in
relation to white students) causes African American
students to "disconnect" their performance in relation to
12


self-esteem. This disconnection, Steele concludes, is
consistent with the disidentification theory (p. 624).
Hare and Costenell (1985) posited that African
American students disidentified with school and home life
because they knew that their chances of performing well
in school and being accepted in a positive manner at home
were not good. They identified with peer-group relations
because they were accepted by their peers and experienced
successful relationships.
Using the Rosenberg's Self-View Inventory,
Osborne's (1995) second prediction (that African American
students tend not to report lower self-esteem than white
students) also was supported. Students' responses to the
survey showed that African American students' self-esteem
was significantly higher than white students (p. 453).
Osborne's third prediction (that correlations
between measures of academic achievement and global
self-esteem should be moderate and significant for both
white and African American students early in their
education, but that over time the correlations should
weaken for African Americans as they disidentify) was
only partially supported. Over the two-year period of
the study, in eighth grade both African American students
and white students showed significant correlations
13


between self-esteem and academic achievement. In tenth
grade, the correlation between self-esteem and academic
achievement dropped for African American male and female
students, increased for white male students, and remained
somewhat unchanged for white female students (p. 452).
Results support Steele's (1992) disidentification theory.
As African American students progressed in school
(specifically from eighth grade to tenth grade) students
started the disidentification process. Osborne suggests
"It is probable that African American students begin very
identified with school and remain that way until sometime
after eighth grade. At that time they may begin to see
the academic environment as discriminatory and lacking in
rewards, and begin disidentifying" (p. 453). To determine
self-esteem Osborne used seven items from Rosenberg's
Self-View Inventory. These items rated global
self-esteem. No items were specific to academic
self-concept.
Problem Focus
This study focuses on the self-esteem, academic
self-concept, and academic achievement of fifth-grade,
seventh-grade, and tenth-grade African American students
in a predominantly white, middle-class, and suburban
school district. These three grades were chosen because
14


the school district annually tests all students in these
grades using the Iowa Test of Basic Skills (ITBS).
Studying these three grades across elementary, middle,
and high school using a standardized measure, students'
semester report card grades, and interviews will enhance
opportunities for study analyses that involve young
adolescents and teenage students.
African American students in a predominantly white
middle-class school district were chosen because this
group has not been the subject of much research. African
American students' academic achievement is widely
discussed, yet little is known about African American
students from middle-class backgrounds.
Accountability reports in this middle-class,
suburban school district reveal that students score above
state and national averages on all areas of the Iowa Test
of Basic Skills (ITBS). African American students in
this district, however, consistently score below the
district, state, and national average of all students.
The Report in Brief of the National Center for
Educational Statistics (NAEP, 1992) reveals that the
trend found in this school district is similar across the
country regardless of students' socioeconomic
backgrounds. African American students' standardized
15


achievement test scores at the state and national levels
are lower than those of their white counterparts.
Investigating this phenomenon in a school district
where academic achievement on standardized tests is above
the state and national level and gathering data only from
African American students may lead to a better
understanding of the relationship between self-esteem,
academic self-concept, and academic achievement. The
specific research question is: What is the relationship
between self-esteem, academic self-concept, and academic
achievement among African American students in a
predominantly white suburban school district?
Hypotheses to examine include: Hypothesis 1: Students'
self-esteem scores will differ according to grade levels.
High school students will have higher self-esteem scores
than will elementary or middle school students. Middle
school students will have higher self-esteem scores than
will elementary students.
Rosenberg (1986) indicates that self-esteem
improves with age. In a review of research on
self-esteem and age, Rosenberg gathered data from
McCarthy and Hoge's (1982) longitudinal study that
revealed that students' self-esteem improved
significantly from grades 7 to 12. Reasons for this
16


improved self-esteem are adolescents' feelings of more
freedom and teens feeling accepted by their peers
(McCarthy & Hoge, 1982, 379). Rosenberg also reviewed
self-esteem and age research conducted by O'Malley and
Bachman (1983) involving adolescents and young adults.
O'Malley and Bachman's research posit that consistent
improvement in self-esteem appears to occur between the
ages of 13 and 23.
Hypothesis 2: The relationship between
self-esteem scores and academic achievement among African
American students will differ according to grade levels.
Elementary students will have a stronger relationship
between self-esteem scores and academic achievement than
will middle school or high school students. Middle
school students will have a stronger relationship between
self-esteem scores and academic achievement than will
high school students.
Steele (1992) introduced the disidentification
theory to explain how African Americans view their
self-esteem in relation to their academic achievement.
Steele contends that African Americans dissociated their
academic achievement from their overall self-esteem.
This disidentif ication causes students to relate their
17


self-esteem to other nonacademic measures such as peer
relations. Osborne (1995) used Steele's
disidentification theory as a conceptual framework and
found that overtime, as African Americans continued in
school, the correlation between self-esteem and academic
achievement weakened. African American students'
self-esteem scores tended not to be related to their
overall academic achievement.
Hypothesis 3: The relationship between academic
self-concept scores and academic achievement among
African American students will differ according to grade
levels. High school students will have a stronger
relationship between academic self-concept scores and
academic achievement them will middle school or
elementary students. Middle school students will have a
stronger relationship between academic self-concept
scores and academic achievement than will elementary
students.
Academic self-concept scores serve to predict how
well students perceive themselves in relation to their
ability in academic subjects (Bums, 1979). Marsh
(1990a) formulated the frame of reference theory to
investigate the relationship between academic
18


self-concept; and academic achievement. Marsh posited
that a student's internal frame of reference in relation
to how well a student believes he or she performs in a
specific class determines the student's academic
self-concept. In Brookover, Erickson, and Joiner's
(1967) longitudinal study, findings indicate that as
students increase in age, a stronger relationship exists
between academic self-concept and academic achievement.
Data will be gathered using Marsh's Self
Description Questionnaire (SDQ), the Iowa Test of Basic
Skills (ITBS) scores, semester report card grade averages
in four subjects (mathematics, reading/language arts,
science and social studies), and interviews. Data
gathered from interviewing African American students may
provide more insight than responses to survey questions
only.
All African American students in grades 5, 7, and
10, during the 1998-1999 school year were contacted to
participate in this research study. Students receiving
parental permission to participate were given Herbert
Marsh's Self Description Questionnaire (SDQ). This
questionnaire was designed to measure self-esteem and
academic self-concept. Students' Iowa Test of Basic
Skills (ITBS) scores and grade card reports are used to
19


measure academic achievement. To gain in-depth
information on the relationship between self-esteem,
academic self-concept, and academic achievement
interviews are conducted with six students (two from each
grade level).
Limitations of the Study
Only African American students participated in this
study. Student participation in this study was
nonrandom. All African American students in grades five,
seven and ten who received parent approval participated
in the study. Six students from this group, two from
each grade level, participated in an in-depth interview.
This study also is limited by the validity of the
instruments used to measure self-esteem, academic
self-concept, and academic achievement.
Definition of Terms
Academic Achievement: student performance in reading,
language arts, mathematics, social studies, and science
as measured by the Iowa Test of Basic Skills (ITBS) and
teacher grade reports.
Self-Esteem/Self-Concept: "The evaluation which the
individual makes and customarily maintains with regard to
himself: it expresses an attitude of approval or
20


disapproval, and indicates the extent to which the
individual believes himself to be capable, significant,
successful, and worthy" (Coopersmith, 1967, p.5).
Academic Self-Concepts Students' perception of
themselves in relation to their ability to learn or
students' perceptions of their abilities in relation to
how they perform in academic subjects (Burns, 1979).
Summary
The remaining chapters are organized in the
following manner: Chapter 2 reviews literature related
to self-esteem. Several definitions from key researchers
are discussed. Studies that focus on the relationship
between self-esteem and academic achievement are
reviewed. In addition, a review of research on the
relationship between academic self-concept and academic
achievement is given. The last portion of chapter 2
reviews literature specific to African American students'
self-esteem and academic achievement. Chapter 3
discusses the instruments and methods used to collect
data on students' self-esteem, academic self-concept and
academic achievement. Chapter 4 analyzes the findings
related to each hypothesis. Chapter 5 provides
conclusions and recommendations.
21


CHAPTER 2
REVIEW OF THE LITERATURE
This literature review focuses on four sections:
(a) the definition of self-esteem and related research,
(b) self-esteem and academic achievement, (c) academic
self-concept and academic achievement, and (d) African
American students' self-esteem and academic achievement.
In the first section, research on the definitions and
concepts of self-esteem is presented. Key researchers
discussed in Mruk's (1995) work on self-esteem are
presented. These researchers added much to self-esteem
literature.
Section two reviews research on self-esteem and
academic achievement. This research focuses on studies
by Bachman and O'Malley (1986), Gwin (1990), Shokraii
(1996), and others. Research on whether self-esteem
programs enhance academic achievement is also presented.
Researchers studying the relationship between self-esteem
and academic achievement generally have used a global
self-esteem instrument to collect data (Bachman &
O'Malley, 1984; Hare, 1979; Madhere, 1991; Maruyama,
22


et al., 1981; Osborne, 1995; Voelkl, 1993; Walker, 1991).
Section three reviews research specific to academic
self-concept and academic achievement. Major works
reviewed in this section are from research done by Marsh
(1990a). Marsh posits that research on self-esteem in
relation to academics must use measurements that
specifically address academic content areas.
The fourth section provides a review of literature
on African American students' self-esteem and academic
achievement. This review focuses on the paradox that
seems to exist among African American students who report
high self-esteem yet whose academic achievement is lower
than their white counterparts. Possible reasons for this
paradox are also reviewed.
Self-Esteem/Self Concept
In order to understand the relationship between
self-esteem and achievement among African American
students, it is important to clarify the concepts of
self-esteem and self-concept. According to Branden
(1990) the cultural definition of self-esteem did not
appear until the late 1980s. Prior to this period
self-esteem research focused primarily on clinical
applications and definitions. This is also when
significant research related to self-esteem appeared.
23


Even then, Branden argues, "There is still no general
consensus about what the term means" (p. 2).
Numerous articles and books have been written on
self-esteem in order to give it definition and meaning.
Mruk (1995) published a scholarly codification of the
various definitions of self-esteem derived by
acknowledged research experts in the major psychological
and sociological works on self-esteem. Mruk reported
that over 6,780 articles and 557 books (from 1974 through
1993) had been written directly relating to self-esteem.
He also reported that an additional 20,000 articles had
been written indirectly relating to self-esteem. He
further reported that over 2,080 self-esteem-related
measurement instruments were then in existence
(p. 10). This section reviews the various definitions
researchers have given to the term self-esteem.
Mruk organized the definition of self-esteem in
several ways. In his review of the authors who studied
self-esteem, several common themes emerged. Authors
studied included James, White, Rosenberg, Coopersmith,
Branden, Epstein, Bednar, and Pope et al. Two of the
self-esteem themes that emerged were competence and
worthiness. Authors defining self-esteem in terms of
competence and worthiness include James, White, and Pope
24


et al. Authors using worthiness in their definition
include Rosenberg, Coopersmith, Epstein, and Bednar.
Branden's definition of self-esteem is the only one that
includes both competence and worthiness (Mruk, 1995, p.
18).
Mruk also grouped the definitions of self-esteem on
the basis of whether they emphasized change or stability.
In this variation, most of the researchers stressed
stability (global self-esteem). According to Mruk, James
tied his definition of self-esteem to dynamic processes
and used a ratio of successes/pretensions to express
this. Mruk indicates that Pope et al. and James also
tied self-esteem to openness. Openness, Mruk asserts,
enables one's self-esteem to change over time. This
allows for self-esteem to be enhanced. Authors defining
self-esteem using the term stability include White,
Rosenberg, Coopersmith, Epstein, and Bednar (p. 20).
The last variation used by Mruk to analyze the
definition of self-esteem was the authors' emphasis on
cognitive terms (evaluation based on either positive or
negative) or affective terms (evaluation based on an
emotional or feeling state) to describe self-esteem.
Rosenberg, Coopersmith, and Pope et al. define
25


self-esteem with emphasis on cognitive terms. James,
White, Branden, Epstein, and Bednar use affective terms
in their self-esteem definitions (p. 19).
From his review of the research, Mruk concludes
that an appropriate definition of self-esteem displays a
continuum characteristic and involves at least three
essential components (a) competence and worthiness and
the relationship between them, (b) stability and
openness, and (c) cognitive and affective processes of a
person's life (p. 21). Competence and worthiness and the
relationship between them creates self-esteem.
Self-esteem exists in the form of a dynamic structure
because only a process or phenomenon can exhibit both the
characteristics of stability and openness. And finally
an appropriate definition of self-esteem must embody the
real life, lived quality of self-esteem for an
individual, individual experiential evaluations of
outcomes based on feelings of analytic or conscious
awareness (cognitive) or intuitive knowledge (affective).
Rosenberg (1965) indicates that self-esteem "is a
positive or negative attitude toward a particular object,
namely, the self" (p. 30). A person with high
self-esteem has a feeling of self-worth and respects
herself or himself. A person with low self-esteem lacks
26


self-respect (p. 31).
According to Branden (1967) self-esteem has two
interrelated aspects: a sense of personal efficacy and a
sense of personal worth. Self-efficacy refers to a
person's ability to produce a desired outcome.
Self-confidence affects the sense of efficacy.
Self-respect affects the sense of personal worth and
gives the individual a sense of value and a sense of
affirmation.
Branden (1990) further states that:
self-esteem is (a) confidence in our ability to
think, confidence in our ability to cope with the
challenges of life, and (b) confidence in our right
to be happy, the feeling of being worthy,
deserving, entitled to assert our needs and wants
and to enjoy the fruits of our efforts, (p. 3)
Coopersmith's (1967) definition of self-esteem
also focuses on a feeling of worthiness. Coopersmith
defines self-esteem as "the evaluation which the
individual makes and customarily maintains with regard to
himself: it expresses an attitude of approval or
disapproval, and indicates the extent to which the
individual believes himself to be capable, significant,
successful, and worthy" (p. 5).
Purkey (1970) characterizes the self as being
organized and dynamic. According to Purkey this
organized self reflects a sense of harmony and
27


orderliness (p. 7). Purkey further asserts that the
human tendency to enhance, protect, and maintain is a
reflection of the dynamic self. Purkey states that
"people are constantly trying to behave in a manner which
is consistent with the way they view themselves" (p. 12).
He further posits that "we express our self-concept with
our behavior" (p. 13). In other words, if academic
achievement has positive outcomes for students they will
achieve and grow in self-esteem.
Gergen (1971) defines self-esteem as a person's
general evaluation of self (p. 50). This is consistent
with Rosenberg's self-worth definition. Rosenberg and
Simmons (1972) also describe self-esteem in terms of
self-worth: "a person with high self-esteem considers
himself a person of worth" (p. 9). Low self-esteem,
according to Rosenberg and Simmons, means a "person
considers himself unworthy, inadequate, or otherwise
seriously deficient as a person" (p. 9).
Shavelson, Hubner, and Stanton (1976) used a
similar definition for self-concept. They state,
"Self-concept is a person's perception of himself. These
perceptions are formed through his experience with his
environment . and are influenced especially by
28


environmental reinforcements and significant others"
(p. 411).
In their research, Marsh, Smith, and Barnes (1983)
define self-concept as "an individual's perception of
self, and is formed through experience with the
environment, interactions with significant others, and
attributions of his or her own behavior" (p. 335).
Rosenberg, cited in Suls and Greenwald (1986),
indicates that self-esteem "primarily involves feelings
of self-acceptance, self-liking, and self-respect
(p. 120). Rosenberg further asserts that feelings of low
self-esteem are associated with depression, unhappiness,
and discouragement.
This literature review provided self-esteem and
self-concept definitions with overlapping terms. Most
authors view self-esteem and self-concept as a person's
feeling of worth based on environmental influences. For
the purpose of this study Coopersmith's (1967) definition
of self-esteem will be used: "The evaluation which the
individual makes and customarily maintains with respect
to himself: it expresses an attitude of approval or
disapproval, and indicates the extent to which the
individual believes himself to be capable, significant,
29


successful, and worthy" (p. 5). In order to examine the
relationship between self-esteem, academic
self-concept and academic achievement among African
American students the perceptions these students have in
relation to whether they believe themselves to be
capable, significant, successful, and worthy will assist
in understanding the achievement paradox that exists.
Related Research
In Rosenberg's (1965) research involving ten high
schools selected at random from public high schools in
New York State, he evaluated how different social
experiences relate to levels of self-esteem (p. 32). The
sample consisted of 5,024 high school juniors and seniors
from the ten high schools. Findings revealed that the
measurement of social class and self-esteem showed those
students who were members of upper and middle class
socioeconomic groups reported higher self-esteem (51
percent and 46 percent respectively) than lower
socioeconomic groups (38 percent). Other social
variables Rosenberg found to be associated with
self-esteem included religion, race, nationality, family
structure, and participation and leadership in the
community (p. 35).
30


In his research, Coopersmith (1967) focused on
conditions that lead people to value themselves as
worthy. To investigate the antecedents, consequences,
and correlates of self-esteem, Coopersmith studied white,
male preadolescents. Four interrelated studies were
carried out. These studies consisted of:
(1) selection of subjects who differed in
self-esteem; (2) the clinical evaluation of
subjects on a battery of ability, projective and
personality questionnaire tests, and a clinical
interview; (3) observation and measurement of the
subject's behavior in a series of laboratory
experiments that were theoretically related to
self-esteem; and (4) ascertaining the antecedents
of self-esteem by interviews and questionnaires
administered to the mother of the subject and to
the subject, (p. 8)
Initial data were gathered from 1,748 subjects.
From this pool, Coopersmith chose 85 subjects who
differed in self-esteem to participate in the rest of the
research. Findings revealed a positive relation between
social class and self-esteem. Higher social class
individuals were more likely to have high self-esteem.
This positive relation between social class and
self-esteem supports previous research conducted by
Rosenberg (1965).
On the other hand, Coopersmith found that no
relationship existed between a child's self-esteem and
whether or not the mother was employed or unemployed.
31


Sons with high self-esteem had mothers who were more
self-reliant and fathers who gave them more attention and
appeared to be more concerned about them.
Research has also been conducted to determine if
self-esteem improves over time. Rosenberg, cited in Suls
and Greenwald (1986), posits that changes in influences
determining how a person evaluates her or his self-esteem
occur as children increase in age (p. 123). Fifth-grade
students may decide their self-esteem levels based on how
well they play baseball or excel in a particular sport.
They may equate self-esteem with how their parents
perceive them. A tenth-grade student may decide her or
his self-esteem level based on peer acceptance.
McCarthy and Hoge's (1982) longitudinal study of
middle childhood, adolescence, and adult self-esteem
revealed that self-esteem improved significantly between
grades 7 and 12. McCarthy and Hoge suggest several
reasons for this improved self-esteem. People with high
self-esteem will work to maintain it. People with low
self-esteem will work to improve it, thus, creating a
higher average over time. Adolescents' enhanced
interpersonal skills may account for the improved
32


self-esteem as well. Adolescents' feelings of more
freedom and acceptance from their peers may also account
for the improved self-esteem (p. 379).
Self-Esteem and Academic Achievement
In the previous section, definitions of
self-esteem and related research were reviewed. This
section reviews the literature on the relationship
between self-esteem and academic achievement. Also
reviewed are research studies that examine whether
self-esteem programs enhance academic achievement.
Self-esteem and its relation to academic
achievement are central themes of numerous research
studies (Bachman & O'Malley, 1984; Crocker & Major, 1989;
Osborne, 1995; Purkey, 1970; Steele, 1992; Voelkl, 1993).
In his book Self Concept and School Achievement. Purkey
(1970) begins Chapter 1 by retelling a story from Lowry's
(1961) opening remarks given in a conference presentation
entitled "The Mouse and Henry Carson." In Purkey's
account of Lowry's speech, Purkey tells how a mouse got
into the Educational Testing Service office and triggered
an apparatus as the apparatus was scoring Henry Carson's
College Entrance Examination. Apparently, Henry was an
average high school student, unsure of himself and his
abilities (p. 1). The mouse changed all this for Henry.
33


Perhaps Henry's scores would have been average had it not
been for the mouse. However, what emerged from the
computer on Henry's score sheet were 800s in both the
verbal and quantitative areas. Naturally, when Henry's
school learned of his scores, the word spread rapidly
that he was a gifted student. Teachers reevaluated their
beliefs about Henry's abilities, counselors viewed Henry
differently and colleges began recruiting Henry.
As positive feelings about Henry and his abilities
increased, Henry became a better student. A
self-fulfilling prophecy occurred. Henry achieved
greatness because he believed in himself and thought
others believed in him. Purkey uses this story to make
two points: "The way in which a student views himself
and his world are (a) products of how others see him, and
(b) primary forces in his academic achievement" (Purkey,
1970, p. 2).
Studies to determine whether programs to increase
students' self-esteem will lead to increased student
achievement have been conducted. Schools are offering
courses on self-esteem in hopes of enhancing academic
achievement (Gwin, 1990; Shokraii, 1996). Schools are
grappling with the question "which comes first,
achievement or self-esteem" (Shokraii, 1996, p. 1). Gwin
34


(1990) argues that, misguided self-esteem efforts are
keeping some students from doing their best. Gwin
indicates that we are patting students on the back to
increase their self-esteem without holding them
accountable for work completion and high levels of
performance. Gwin maintains that these efforts to
increase self-esteem without holding our students to high
academic standards are hurting our students. Gwin asks,
"Are teachers using precious teaching time to make a
child feel good about himself only to see this same child
perform poorly on national exams? Is thought being given
to the fact that most children automatically feel good if
they improve academically" (p. 16)?
Shokraii (1996) argues that "Compelling research
from around the world lends empirical proof to the
traditional claim that achievement precedes self-esteem.
There is, in fact, almost no link between low self-esteem
and any number of social pathologies, including poor
school performance, drug abuse, and teenage pregnancy"
(p. 1). Shokraii discusses two types of self-esteem:
earned self-esteem and global self-esteem. Earned
self-esteem is the self-esteem people earn by way of
their accomplishments. Global self-esteem refers to a
general sense of pride in oneself. Global self-esteem is
35


not grounded in a particular skill or accomplishment.
Shokraii argues that this kind of self-esteem at best is
meaningless and at worst is harmful (p. 3). Global
self-esteem, according to Shokraii, gives students a
false sense of accomplishment. It leads to empty praise
and may cause students to desire praise without putting
forth effort and hard work in order to accomplish a task.
Shokraii further posits that by granting praise to
African American students without the evidence of
accomplishment may lead to an exaggerated sense of
accomplishment and achievement.
Researchers, according to Shokraii, contend that
global self-esteem can inspire academic success.
Shokraii, on the other hand, posits that the opposite is
true. Academic success should inspire global
self-esteem. Increased academic achievement should cause
students to believe in their self worth and motivate them
to want to achieve at even higher levels thus increasing
their self-esteem.
Researchers have examined whether interventions to
enhance self-esteem would correlate to increased academic
achievement. Scheirer and Kraut (1979), for example,
analyzed research studies that used self-esteem
intervention measures to increase achievement. They
36


grouped their review into several categories: preschool
children, primary grade school children, junior high
school, and high school children. Reviewing the research
studies, they chose only those meeting certain criteria:
(a) a control group for comparison with the treatment
group had to be present, (b) measures of both
self-concept and academic achievement had to be included
after an intervention program that was intended to
increase academic achievement, and (c) pre and post-
intervention measures of self-concept and academic
achievement had to be included (p. 134).
Scheirer and Kraut found that the studies indicated
no causal connection between self-concept and academic
achievement. In each study, the treatment group did not
show any significant increases in achievement when
measures to enhance self-concept were used. The authors
caution educators not to assume that enhancing a person's
self-concept or self-esteem will automatically lead to
increased achievement.
in spite of the findings presented by Scheirer and
Kraut (1979), programs to enhance self-esteem in order to
improve student achievement are on going. Walker (1991)
discusses the results of a study on the effects of a
counseling program on the self-esteem and achievement of
37


300 elementary students in grades 3 through 8. Based on
the students' previous academic records they were
considered "at risk." Findings, however, revealed
results similar to those of Scheirer and Kraut (1979).
The self-esteem of the individuals who participated in
the counseling programs was higher at the end of the
school year. This did not, however, translate into
increased academic achievement. No significant
differences existed in the academic performance between
students in the counseling group and those in the control
group.
Findings from other research that examined the
relationship between self-esteem and academic achievement
suggest that it is students' academic perceptions and
their actual academic abilities that determine their
achievement rather than their levels of self-esteem.
Bohrnstedt and Felson (1983) examined the relationship
between children's achievement and self-esteem using
three different types of achievement areas: academics,
sports, and popularity. Data were collected from 415
sixth through eighth grade boys and girls. Self-esteem
was measured using a self-esteem questionnaire. Grades
were determined using grade-point averages in four
subjects. Athletic performance was based on a basketball
38


test:. Popularity information was obtained by asking all
students to name the three boys and three girls whom they
liked best in their fourth-period class. Perceptions of
academic ability were measured by asking students how
smart they thought they were (p. 47). Findings indicate
that children's self-esteem affected how popular they
thought they were. For academics and sports, perceptions
of achievement were strongly related to actual
achievement, therefore, affecting their self-esteem
rather than the reverse (p. 43).
Bachman and O'Malley (1986) examined whether school
academic climates had any impact on self-concepts of
academic ability, global self-esteem, and long-range
educational attainment. Data were gathered from 1,487
men from the Youth in Transition high school study.
Findings indicate that, after controlling for
socioeconomic status
There were only small negative effects of school
mean ability on self-concepts and self-esteem.
Educational attainment five years beyond high
school was strongly influenced by background,
ability, and grades . there was little
additional impact from self-concepts and self-
esteem and no overall effect attributable to school
climate. (p. 35)
School mean ability referred to the ability level of
students in the class computed as a mean ability score
determined by student responses on three instruments
39


(Ammons and Ammons Quick Test, -the Gates Test of Reading
Comprehension and the General Aptitude Test Battery).
Bachman and O'Malley concluded that it is the
actual abilities students demonstrate, not the
self-concept or self-esteem, that makes a difference in
educational attainment. They confirmed this
longitudinally by studying the same men from age 15 to
23. They do suggest, however, that it may be possible
that earlier years may have played a more crucial part in
shaping later outcomes (p. 45). They contend that, for
long-range educational attainment, it is actual ability
that matters not self-concept.
Maruyama et al., (1981) examined the relationship
between social class, ability, educational achievement,
and self-esteem. Data were collected from a total of 715
children ranging in age from 9 to 15 years and born
between 1961 through 1964 (p. 962). Findings revealed
that social class and ability were strongly related to
one another and to measures of achievement and
self-esteem (p. 967). They found no direct relation
between achievement and self-esteem. Achievement and
self-esteem were related only to the extent that they
shared the common variables of social class and ability
as the common causes (p. 973). Results indicated that
40


self-esteem does not cause achievement. Further, no
strong evidence indicated whether achievement caused
self-esteem.
According to Shavelson and Bolus (1982), research
on the relationship between self-concept and academic
achievement using grades and test scores to measure
achievement produce positive correlations. Shavelson and
Bolus posit that self-concept is multifaceted and
hierarchical. They further argue that measures of
academic self-concept and achievement tend to have higher
correlations than correlations between global
self-concept and academic achievement (p. 6).
Marsh (1992) posits that in order to gain a better
understanding of the relationship between self-esteem and
academic achievement a measure specific to academic
content areas must be included. Expanding on Shavelson
and Bolus's (1982) hierarchical model of self-concept,
Marsh designed the Self Description Questionnaire (SDQ).
The SDQ was developed to measure seven facets of
self-concept (physical abilities, appearance, peer
relations, parent relations, reading, mathematics, and
general school). The general self subscale was later
added (Marsh, 1990b).
41


This study uses Marsh's SDQ to examine the
relationship between self-esteem, academic self-concept,
and academic achievement. Marsh's SDQ was chosen because
it ask questions specific to global self-esteem, academic
self-concept and specific to the content areas of
mathematics and reading.
Academic Self-Concept and
Academic Achievement
Academic self-concept as defined by Burns (1979),
refers to students' perceptions of themselves in relation
to their ability to learn or students' perceptions of
their abilities in relation to how they perform in
academic subjects. This section of the literature review
focuses on research that examines the relation between
academic self-concept and achievement.
Marsh (1990a) and others (Marsh, Byrne &
Shavelson, 1988; Marsh & O'Neill, 1984; Williams &
Montgomery, 1994) indicate that a more positive
correlation will exist between self-esteem and academic
self-concept if a measure specific to academic
self-concept is correlated with specific academic content
subjects. Marsh (1990a) formulated the frame of
reference theory to investigate the relationship between
academic self-concept and academic achievement. This
42


model suggests that both internal and external frames of
references determine a student's academic self-concept.
A student's internal frame of reference refers to the
student comparing her or his academic achievement in
relation to a specific subject. The student believes
that he or she is better in English than in mathematics.
An external reference for the student is making
comparisons of how well he or she performs in English
compared to other students in this particular subject.
Studying the relationship between academic
self-concept and academic achievement. Marsh (1992) used
14 components of academic self-concept specific to
8 school subjects (p. 35). The sample consisted of 507
high school boys. Results indicate that academic
self-concept related more strongly to academic
achievement than general measures of self-concept.
Findings revealed that "relations between academic
self-concepts and academic achievements are more content
specific than previously assumed" (p. 35).
Williams and Montgomery (1994) examined the
academic self-concept of academically capable students
using Marsh's (1990a) frame of reference theory. The
sample included 103 students enrolled in honors courses.
Findings indicated that the frame of reference theory was
43


validated with academically capable students. Students'
academic self-concept was believed to be influenced by
both an internal and external comparison of performance.
A significant relationship existed between students'
academic self-concept and their academic achievement.
Brookover et al. (1967) studied the relationship
between academic self-concept and academic achievement of
307 girls and 255 boys between the ages of 12 and 17.
Their longitudinal study extended for six years.
Findings indicate that a significant relationship existed
between academic self-concept and academic achievement at
each age level. A stronger correlation between academic
self-concept and academic achievement occurred as
students moved from one grade to the next. Any changes in
academic achievement resulted in changes of academic
self-concept. Brookover et al. concluded that academic
self-concept gives a higher correlation with academic
achievement than general measures of self-concept.
Research conducted by Mboya (1989) produced similar
results. Mboya studied the relationship between
self-concept of academic ability and academic achievement
to determine if this relationship correlated more
strongly than the relationship between global
44


self-concept and academic achievement (p. 39). Data were
collected from 229 tenth-grade students. Results
revealed a positive relationship between global
self-concept and academic achievement and between
self-concept of academic ability and academic
achievement. However, the relationship between
self-concept of academic ability and academic achievement
correlated more strongly. According to Mboya, a
student's self-concept of academic ability is significant
in helping the student make decisions related to school
success. Mboya posited that students learn what they
perceive they are able to learn (p. 27).
Research investigating the relationship of
self-esteem to academic achievement reveals
that the correlation between academic achievement
and self-esteem rises significantly when a
domain-specific self-concept, such as academic
self-concept was considered (Watkins & Gutierrez, 1989).
Based on research (Brookover et al., 1967; Marsh, 1990a,
1992; Mboya, 1989; Williams & Montgomery 1994)
investigations of the relationship between academic
self-concept and academic achievement show a more
positive correlation than just investigating the
relationship between global self-esteem and achievement.
45


Self-Esteem and African American
Student. Achievement
A number of researchers contend that
self-esteem should relate to academic achievement
(Gergen, 1971; Purkey, 1970; Rosenberg, 1965). The
literature on African American self-esteem reports
contradictions. Some research studies investigating
African American students' self-esteem and academic
achievement show a positive correlation. In other words,
students who perceive that they can achieve at high
levels and do well in school actually do achieve at high
levels. These research studies examined global
self-esteem in relation to academic achievement of
African American students.
Research from the early 1960s showed African
American students' self-esteem to be lower than whites
(Drury, 1980). One reason for this, Drury contends, is
that early studies were hampered by the use of small
sample groups (usually nursery and kindergarten students)
and because they relied on inferential measures of
self-esteem (p. 89). However, in Drury's review of
contemporary literature that employed more substantial
samples and provided more direct measures of self-esteem,
15 of the 24 studies indicated African American students'
46


self-esteem was as high or higher than white students.
In five of the 24 studies, white students' self-esteem
was higher, and in four comparisons no significant
differences were found.
Voelkl (1993) established that differences in
academic achievement between minority students and white
students have been studied as early as 1920. She reports
that findings from these studies consistently reveal
that, "on average, African American students score below
white students on tests of academic achievement" (p. 42).
In spite of the apparent lower academic achievement among
African American students compared to white students,
studies indicate that African American students do not
report lower self-esteem (Bachman & O'Malley, 1984;
Crocker & Major, 1989; Osborne, 1995; Steele, 1992;
Voelkl, 1993).
Hare (1979) compared the achievement and
self-perception of 5th grade black girls as compared to
black boys and white girls. Findings indicate black
girls share with black boys a lower achievement
orientation, a higher self-concept of ability, and
significantly lower reading and math scores compared to
white students (p. 20). This higher self-concept, lower
achievement finding is consistent with results from other
47


research studies (Bachman & O'Malley, 1984; Hare &
Costenell, 1985; Madhere 1991; Osborne, 1995; Rosenberg &
Simmons, 1972; Voelkl, 1993).
Simmons, et al., (1978) measured the self-esteem
and academic achievement of 798 African American and
white students in grade 6 and as they moved to grade 7.
Findings indicate that (a) African American students
appear to have higher, rather than lower, self-esteem
than white students, (b) girls of both races have lower
self-esteem than boys (white girls exhibited the lowest
self-esteem), and (c) African American students from
broken homes reported lower self-esteem in desegregated
than in segregated schools (p. 86). Research that
follows examine why this phenomenon (African American
students' high self-esteem, low achievement in relation
to white students) exists.
According to Osborne (1995), Steele's (1992) theory
of disidentification accounts for both African American
students' poor academic performance and their
paradoxically high self-esteem. Steele argues that
"Black students must have acquired an imperviousness to
poor school performance" (Steele, 1992, p. 623).
Steele (1997) contends that African American
students' ability stigmatization (e.g., poor performance
48


in relation to white students') cause African American
students to "disconnect" their performance in relation to
self-esteem. This disconnection, Steele concludes, is
consistent with the disidentification theory (p. 624).
Stereotype threat is another phenomenon that might
interfere with African American's academic achievement in
relation to their white counterparts. Steele and Aronson
(1995) define stereotype threat as "being at risk of
confirming, as self-characteristic, a negative stereotype
about one's group" (p. 797). Steele and Aronson (1995)
argue that:
Whenever African American students perform an
explicitly scholastic or intellectual task, they
face the threat of confirming or being judged by a
negative societal stereotypea suspicionabout
their group's intellectual ability and competence.
This threat is not borne by people not stereotyped
in this way. And the self-threat it causes
through a variety of mechanismsmay interfere with
the intellectual functioning of these students,
particularly during standardized tests.
(p. 797)
Testing this hypothesis, Steele and Aronson found that
African American students underperformed in relation to
white students in the ability-diagnostic conditions but
not in the nondiagnostic condition. The ability-
diagnostic conditions were established using scholastic
and intellectual tasks. When these same tasks were
presented in nondiagnostic conditions (e.g., not tied to
49


scholastic or intelligence) African American students
performed as well or better than white students.
African American student achievement is widely
reported as being lower than that of their white
counterparts. According to Steele and Aronson this
stereotype threat (African American students perceiving
their abilities in a negative manner compared to white
students) causes African American students to
underperform.
Steele (1999) contends that this stereotype threat
persists even among African American students from middle
class backgrounds. Steele argues that there appears to
be something beyond class, something that is racial
(stereotype threat) that is depressing the academic
performance of African American students (p. 46). This
stereotype threat, Steele posits, causes
disidentification among African American students.
Rosenberg and Simmons (1972) interviewed over 1900
students in grades 3 to 12 about self-esteem and
achievement. They found that black students'
self-esteem was as high as white students even though
black students' achievement was lower. Even when
socioeconomic status was controlled, black students'
50


self-esteem remained high. Rosenberg and Simmons argue
that school performance should be relevant to
self-esteem. They contend that if black students have
lower grades them white students their self-esteem should
be lower than white students. Their research results
revealed the opposite.
They reasoned that black students low academic
achievement and high self-esteem may be due to black
students' coping and defense mechanisms, and black
students determining their self-esteem on measures other
than how well they do in school. In their review of
twelve studies between 1963 and 1970, Rosenberg and
Simmons found that some of the studies show black
students with lower self-esteem, some with higher
self-esteem, and some showed virtually no difference
(p. 8).
In an attempt to understand why African American
students' academic achievement is lower than their white
counterparts, Ogbu (1978) cites the cultural deprivation
theory as one explanation given in the 1960s. This
theory posits that students from an environment that does
not provide adequate stimulation for normal development
will report lower academic achievement. Ogbu stated that
this theory did not specifically identify African
51


American students. Due to African American students'
coming from environments that did not provide them with
adequate stimulation for their linguistic, cognitive and
social development they accounted for seventy-five
percent of those identified as culturally deprived
(p. 44).
Ogbu (1978) also used the cultural conflict theory
to suggest why African American students' academic
achievement was below their white counterparts. This
theory asserts that African American students "acquire
values, attitudes, and learning styles within their
culture that are different from and in conflict with
those required for success in the public schools and in
wider society" (p. 47). This theory claims that African
American families do not equip their students with the
skills of middle-class students. These skills, the
theory claims, are needed for academic success.
Ogbu, in Gibson and Ogbu (1991) attributes African
American students' low school performance in relation to
white students to African American students being an
involuntary minority group. Because African Americans
were forced to live in America, a caste status was
created. This status, argues Ogbu, resulted in school
experiences that were different for African American
52


students (e.g., inadequate school materials, inadequate
facilities, and poor teacher preparation) thus leading to
lower school performance. African American students,
Ogbu contends, knew that their mobility status, even with
advanced schooling, was limited.
Fordham and Ogbu (1986) suggest that "the fear of
being accused of 'acting white' causes a social and
psychological situation which diminishes black students'
academic effort and thus leads to underachievement"
(p. 176). One reason for this, Fordham and Ogbu claim is
that white America refused to accept the fact that
African Americans are intellectually capable. This led
to African Americans doubting their own intelligence.
Subsequently, African Americans defined academic success
as a white person's privilege and were discouraged from
succeeding academically and discouraged other African
Americans from "acting white" (p. 177).
Fordham (1996) charges in her ethnographic analysis
of African American students at Capital High that
"students' understanding of the ways they are represented
and tortured by the so-called scientific analyses of the
racialized Other affect their academic performance" (p.
11). Students' refusal to assimilate or pass as an Other
(the dominant community and culture) are viewed as
53


academic failures. Fordham argues that this resistance
to assimilate negatively impacts African American
students' academic performance.
In spite of African American students' low
achievement in relation to their white counterparts,
many believe that the performance gap can be eliminated.
Jencks and Phillips (1998) compiled research from
several different authors to examine the black-white
achievement gap. From their studies they conclude that
the black-white test score gap does not appear to
be an inevitable fact of nature. It is true that
the gap shrinks only a little when black and white
children attend the same schools. It is also true
that the gap shrinks only a little when black and
white families have the same amount of schooling,
the same income, and the same wealth. But despite
endless speculation, no one has found genetic
evidence indicating that blacks have less innate
intellectual ability than whites. Thus while it is
clear that eliminating the test score gap would
require more than one generation, we believe it can
be done. (p. 2)
Research studies reveal that the academic
achievement of African American students remains
significantly below the achievement of white students.
Determining if this achievement relates to self-esteem
issues remains a significant reason for research to
continue.
54


CHAPTER 3
METHODOLOGY
The purpose of this study is to examine the
relationship between self-esteem, academic self-concept,
and academic achievement among African American students
in grades 5, 7, and 10. Specifically, the study
addresses the question: What is the relationship between
self-esteem, academic self-concept, and academic
achievement among African American students in a
predominantly white suburban school district? This study
has two aspects, each of equal importance. First, this
is a correlational study to examine the relationship
between academic achievement, self-esteem, and academic
self-concept at three grade levels. Second, this is an
interview study to help explain the relationship between
academic achievement, self-esteem, and academic
self-concept from the perspective of the students
themselves.
This correlational study examines three hypotheses:
Hypothesis 1: Students' self-esteem scores will differ
according to grade levels. High school students will
55


have higher self-esteem scores than will elementary or
middle school students. Middle school students will have
higher self-esteem scores than will elementary students.
Hypothesis 2: The relationship between self-esteem
scores and academic achievement among African American
students will differ according to grade levels.
Elementary students will have a stronger relationship
between self-esteem scores and academic achievement than
will middle school or high school students. Middle
school students will have a stronger relationship between
self-esteem scores and academic achievement than will
high school students.
Hypothesis 3: The relationship between academic
self-concept scores and academic achievement among
African American students will differ according to grade
levels. High school students will have a stronger
relationship between academic self-concept scores and
academic achievement than will middle school or
elementary students. Middle school students will have a
stronger relationship between academic self-concept
scores and academic achievement than will elementary
students.
To test these hypotheses, a correlational study was
conducted to examine the relationship between
56


self-esteem, academic self-concept., and academic
achievement among African American students at three
grade levels. Six students, two from each level, were
interviewed. At each grade level one student reporting
high self-esteem and receiving low grades and one student
reporting low self-esteem but receiving high grades were
interviewed. The purpose of the interview was to gain
in-depth information on how students perceive their
self-esteem, academic self-concept, and academic
achievement in relation to the scores each received on
the SDQ, ITBS, and report card grades for the semester.
Specifically, students were asked questions related to
the three hypotheses to determine if students' answers
given on the SDQ correlated with responses given during
the interview.
Sample
A suburban school district with less than a ten
percent African American population was the focus of this
study. All African American students in grades 5, 7, and
10 during the 1998-1999 school year were contacted to
acquire permission to give Herbert Marsh's Self
Description Questionnaire (SDQ). The African American
student population in grades 5, 7, and 10 consisted of 50
students in grade five, 42 students in grade seven, and
57


32 students in grade 10. Six students (two from each
grade level) gave in-depth interviews. Data was
collected from four sources: the SDQ, Iowa Test of Basic
Skills scores, students' semester report card grade
averages in four subjects (mathematics, reading/language
arts, science and social studies), and individual
interviews with a sample of students.
Data gathered from interviewing African American
students provided in-depth information regarding the
relationship between self-esteem, academic
self-concept, and academic achievement. Most of the
previous research (involving academic achievement and
self-esteem) used global self-esteem survey results and
standardized test scores to collect and analyze data
(Bachman & O'Malley, 1984; Crocker & Major, 1989;
Osborne, 1995; Simmons et al. 1978).
Data Collection
Three types of data were collected: indicators of
self-esteem and academic self-concept as measured by the
Self Description Questionnaire, indicators of achievement
as measured by the Iowa Test of Basic Skills and semester
report card grade averages in four academic subjects, and
structured interviews with a subset of students.
58


The Self Description Questionnaire
The Self Description Questionnaire (SDQ) was used
to determine students' self-esteem and academic
self-concept, it is a multidimensional instrument
designed to measure students' self-concept in academic,
nonacademic and global domains. The Self Description
Questionnaire (SDQ) was originally developed to measure
self-concept in seven areas: physical abilities,
appearance, peer relations, parent relations, reading,
mathematics, and general school (Marsh & O'Neill, 1984,
p. 155). These dimensions of self-concept were first
hypothesized in Shavelson's hierarchical model (Marsh,
Relich, & Smith, 1983, p. 173).
Marsh (1990b) indicates that the general self scale
was later included and was derived from the Rosenberg
Self-Esteem scale (p. 1). Marsh has since designed
several other self-concept instruments. These include
the revised Self Description Questionnaire, now called
the SDQI and contains 76 items, the Self Description
Questionnaire II, and the Self Description Questionnaire
III. The SDQ was originally designed to use with young
students in grades 4 through 6. Marsh (1990b) indicates
that the SDQ is suitable for younger students in grade 2
and even for older students through high school and even
59


college (p. 3). The SDQ instruments (I, II, and III) may
be ordered from The Self Research Centre, University of
western Sydney, MacArthur, P. 0. Box 555, Campbelltown
NSW 2560 Australia.
According to Marsh, Smith, and Barnes (1983),
coefficient alphas for each of the dimensions of
self-concept range from .80 to .92 (p. 339). The
coefficient alphas were obtained after they conducted a
study using the SDQ on 654 fifth and sixth grade
students. The students were asked to complete the SDQ
along with a standardized reading achievement test and
the Intellectual Achievement Responsibility (IAR)
instrument. Results of this study indicate that the
factor analysis of students' ratings demonstrate the SDQ
measures distinct aspects of self-concept. Findings also
provide strong support for the multidimensionality of
self-concept.
This research uses the SDQ designed by Marsh that
contains 56 items. It is found in Appendix A and
contains scales to measure self-concept in seven areas
physical abilities, physical appearance, peer relations,
reading, mathematics, general school and general self
where students are asked to respond with one of six
60


responses (false, mostly false, sometimes false,
sometimes true, mostly true, and true).
Although the entire 56 item SDQ was administered to
students, only general school, general self, reading, and
mathematics subscales are used in this study. General
school mean scores are calculated using questions 2, 7,
13, 22, 27, 40, 46, and 52. Two of the general school
items ares "I'm good at all school subjects," and "I
learn things quickly in all school subjects."
General self mean scores are calculated using
questions 21, 32, 38, 48, 51, 53, 55, and 56. Two
general self items are: "In general I like being the way
I am," and "Other people think I am a good person."
Reading mean scores cure calculated using questions
4, 9, 14, 18, 29, 35, 42, and 54. Two reading questions
are: "I like reading/language arts," and "I learn things
quickly in reading/language arts." Language arts is
included because secondary students taking the SDQ being
enrolled in language arts rather than reading.
Mathematics mean scores are calculated using
questions 10, 15, 19, 24, 30, 36, 43, and 49. Two
mathematics questions are: "I enjoy doing work in
mathematics," and "I am good at mathematics."
61


Marsh (1990b) indicates that in a normative sample
of 3,562 students the internal consistency coefficients
of the SDQ range from .80 to 92. For the subscales used
in this study Marsh determined the coefficient alphas to
be: general school .86, general self .81, reading .89,
and mathematics .89 (p. 45). Marsh concludes that these
results "demonstrate that responses to the SDQ-I reliably
measure facets of self-concept that axe internally
consistent and clearly distinguishable" (p. 45).
African American students in grades 5, 7, and 10
completed the SDQ. Students indicated their grade level,
gender, and age on the SDQ. Student identification
numbers were used instead of studentsr names. Completion
time was approximately ten minutes.
Iowa Test of Basic Skills lITBS)
The Iowa Test of Basic Skills (ITBS) and students'
semester report card grade averages in four subjects were
used to measure students' academic achievement. The
school district in this study used the ITBS in grades 5
and 7. Students in grade 10 were given the Iowa Tests of
Achievement and Proficiency (TAP).
The ITBS and TAP provide data to assist school
districts in determining how their students compare to
students in the same norming groups. Test reliability
62


refers to the stability and consistency of results over
time. Test reliability for the ITBS is consistently
high. According to Riverside 2000 (1994), the fall and
spring test reliabilities for the ITBS in grades 3-8
average .86 and .87, respectively. Core Total and
Composite average values for both fall and spring are .98
(p. 75). The fall and spring average test reliabilities
for the TAP in grades 9-12 are .89 and .90 respectively.
Average Core Total and Composite reliabilities for the
TAP are .98 for fall and spring (p. 91).
The TAP given to students in grade 10 tested
students' basic skills in reading comprehension, written
expression, mathematics, social studies, science and
information processing. The Complete and Core Battery
given to students in grade 5 tested basic skills in the
following areas: reading and reading comprehension,
language, mathematics, social studies, science, and
sources of information.
Students in grade 7 were given the ITBS Survey
Battery. The fall and spring average test reliabilities
for this test are .85 and .86, respectively. The
corresponding Survey Total average for both fall and
spring is .96 (p. 99). The Survey Battery given to
63


students in grade 7 tested basic skills in the following
areas: reading, language, and mathematics.
Riverside 2000 (1994) provides details of test
fairness measures to ensure tests are fair, reliable and
valid to all ethnic and gender groups. The test
development process used helped to ensure test fairness.
Over 100,000 students from schools in 30 states plus Guam
and the Virgin Islands were given tryout versions of
either the Iowa Test of Basic Skills (ITBS), the Tests of
Achievement and Proficiency (TAP), or the Iowa Tests of
Educational Development (ITED). The testing company
deliberately over-sampled minority students to ensure
adequate sample sizes. Educators representing a variety
of racial and ethnic groups reviewed each test item.
Data from the test tryouts and reviewer comments were
analyzed in order to select questions for the final
tests.
According to Riverside 2000 (1994), the ITBS was
first developed in 1935 at the University of Iowa. The
original test was called Iowa Every Pupil Test of Basic
Skills (IEPTBS). The test was renamed in the 1950s.
when authors Lindquist, Greene, Horn, McBroom and Spitzer
developed the IEPTBS they termed basic skills to mean the
range of skills students needed to progress successfully
64


in school (pp. 14-15). In the late 1960s Dr. Hieronymus
and Dr. Hoover developed the modem mathematics
supplement to the ITBS. In the 1980s Dr. Hoover became
the principal author of the ITBS (p. 15).
Students' Grade Reports
Students in each of the grade levels studied
receive semester grades. For the purpose of this study,
students' semester report card grade averages in four
subjects (mathematics, reading/language arts, science,
and social studies) from the 1998-1999 school year are
used. Students' semester report card grades reflect
grades teachers give on assignments from classroom
quizzes, research reports, chapter tests in each subject,
and graded homework assignments. These grades
collectively determine the overall achievement
demonstrated by each student. While no formal validity
and reliability data are given on grade card reports,
these grades are significant in determining how well
students have demonstrated their understanding of subject
matter taught. Students' daily academic achievement in
each subject is determined by the methods each teacher
use to assess academic achievement.
65


Student Interviews
One-on-one interviews were conducted with six
students, two from each grade level, to gather in-depth
information on how students perceive their self-esteem,
academic self-concept, and academic achievement in
relation to the scores each received on the SDQ, ITBS,
and report card grades. Studies have found that by high
school there is little correlation between the academic
achievement of African American students and their own
self-assessment. In particular, African American
students report a high assessment of their self-worth and
ability, yet achieve at lower levels (Osborne, 1995;
Steele, 1992; Voelkl, 1993).
The purpose of the interviews was to explore this
lack of a positive relationship between achievement and
self-concept by conducting in-depth interviews with a set
of students who demonstrated a known discrepancy between
achievement and self-concept. Particularly, at each
grade level one student reporting high self-esteem and
receiving low grades and one student reporting low
self-esteem but receiving high grades were interviewed.
Self-esteem mean scores as determined by responses
to the SDQ and academic achievement mean scores as
determined by students' grade point averages in four
66


subjects were used to determine which students to
interview. The interviews consisted of three students
(one at each grade level) with self-esteem scores above
the mean and academic achievement below the mean and
three students (one at each grade level) with self-esteem
scores below the mean and academic achievement above the
mean.
The interview questions cure found in Appendix B.
The questions focus on three areas: self-esteem, academic
self-concept, and academic achievement. Questions
included are: Do you feel you are able to do things as
well as most people? Explain. Are you satisfied with
your academic achievement? Why/why not? What grades are
you currently getting in reading/language arts and
mathematics? Asking these questions allowed the
researcher to analyze student interview responses in
relation to how they responded to the 56 item SDQ.
Interview data were analyzed by recording (using an
audiocassette player), transcribing, and coding. Coding
was done by selecting categories of student responses.
The categories included general self-esteem, academic
self-concept, general school academic achievement, and
academic abilities in three areas (reading/language arts,
67


mathematics and the ITBS). Student responses were coded
by grade level and by students reporting high
self-esteem and low academic achievement and by those
students reporting low self-esteem and high academic
achievement.
After responses were coded and placed in the
categories mentioned the data were analyzed by reviewing
similarities and differences in responses across grade
levels and across students reporting high self-esteem and
low achievement and those reporting low self-esteem and
high achievement. Interview results cure given in
Chapter 4.
Statistical Analysis
In order to analyze the relationship between
self-esteem and academic achievement and the relationship
between academic self-concept and academic achievement
the SPSS computer data analysis system was used.
Individual student responses to the 56 item SDQ were
entered into an SPSS file. Also included in the file
were students' grade point averages (determined by
averaging semester grades in mathematics,
reading/language arts, science, and social studies),
students' semester mathematics grade, students' semester
reading/language arts grade, students' Total ITBS Reading
68


score, students' Total ITBS Mathematics score, and
students' Total ITBS score.
Mean scores and standard deviation, by grade level,
were computed for students' general self (global
self-esteem), general school (academic self-concept),
mathematics self-concept, and reading self-concept as
measured by the SDQ. Mean scores, by grade level, were
computed for students' academic achievement as measured
by students' overall grade point averages in four
subjects (mathematics, reading/language arts, science,
and social studies). Mean scores, by grade level, were
also computed for students' academic achievement in
mathematics and in reading/language arts.
Pearson product-moment correlations were computed
for the following four areas: (a) global self
(self-esteem) as measured by responses to questions on
the SDQ with academic achievement as measured by grade
point averages and Total ITBS, (b) general school
(academic self-concept) as measured by responses to
questions on the SDQ with academic achievement as
measured by grade point averages and Total ITBS, (c)
reading self-concept as measured by responses to
questions on the SDQ with reading grades as measured by
reading grades on students' grade card report and ITBS
69


total reading scores, and (d) mathematics self-concept as
measured by responses to the SDQ with mathematics grades
as measured by mathematics report card grades and ITBS
total mathematics scores.
Hypothesis 1: Students' self-esteem scores
will differ according to grade levels. In order to
examine this hypothesis general self-esteem mean scores
were computed. Data for this computation were gathered
from students' general self responses on the SDQ.
Hypothesis 2: Elementary students will have a
stronger relationship between self-esteem scores and
academic achievement than will middle school or high
school students. Correlations of self-esteem with
academic achievement were computed using total ITBS scores
and students' grade point averages.
Hypothesis 3: High school students will have a
stronger relationship between their academic self-concept
scores and academic achievement them will middle and
elementary students. Academic self-concept scores were
correlated with students' grade point averages and with
students' total ITBS.
In addition to the above analysis, the SPSS was
used to correlate students' reading self-concept scores
with their semester reading/language arts grades and with
70


their ITBS total reading scores. Students' mathematics
self-concept scores were correlated with their semester
mathematics grades and with their ITBS total mathematics
scores. Results for each hypothesis are reported in
Chapter 4.
71


CHAPTER 4
ANALYSIS OF FINDINGS
The purpose of this study was to examine the
relationship between self-esteem, academic self-concept,
and academic achievement among African American students
in a predominantly white suburban school district.
Hypotheses to examine are given below.
Hypothesis Is Students' self-esteem scores will
differ according to grade levels. High school students
will have higher self-esteem scores than will elementary
or middle school students. Middle school students will
have higher self-esteem scores than will elementary
students. Reviewing self-esteem and age research
Rosenberg (1986) contends that self-esteem improves with
age. McCarthy and Hoge's (1982) longitudinal study
revealed that students' self-esteem improved
significantly for students between grades 7 and 12. One
reason for this is that as students get older their
self-esteem may be based on having more freedom and
feeling more accepted by their peers. O'Malley and
Bachman's (1983) research study involving adolescents and
72


young adults indicates that self-esteem improved
consistently between the ages of 13 and 23.
Hypothesis 2: The relationship between
self-esteem scores and academic achievement among African
American students will differ according to grade levels.
Elementary students will have a stronger relationship
between self-esteem scores and academic achievement than
will middle school or high school students. Middle
school students will have a stronger relationship between
self-esteem scores and academic achievement than will
high school students.
Steele (1992) contends that African American
students dissociate their self-esteem from their academic
achievement. Results of research conducted by Osborne
(1995) show that as African American students continue in
school the correlation between self-esteem and academic
achievement weakens. According to Steele (1992) this
occurs because African American students disidentify
their self-esteem with their academic achievement and
relate their self-esteem to other nonacademic measures.
One such measure is peer relations.
Hypothesis 3: The relationship between academic
self-concept scores and academic achievement among
African American students will differ according to grade
73


levels. High school students will have a stronger
relationship between academic self-concept scores and
academic achievement than will middle school or
elementary students. Middle school students will have a
stronger relationship between academic self-concept
scores and academic achievement than will elementary
students.
Research conducted by Brookover, Erickson, and
Joiner (1967) and Marsh (1990a) indicate that a student's
perception of how well he or she performs in a specific
class determines the student's academic self-concept. As
a student continues in school and the student's grades
improve in a particular subject the student's academic
self-concept in that subject should increase. In their
longitudinal study, Brookover, et al. (1967) found that a
stronger relationship between academic self-concept and
academic achievement existed as students continued in
school.
Steele's theory of disidentification established
the theoretical framework for this study. Steele (1992)
contends that African American students disidentify their
self-esteem with school performance. This
disidentification, according to Steele, accounts for both
74


African American students' poor academic performance and
their paradoxically high self-esteem.
Data were collected from African American students
in grades five, seven, and ten in a suburban school
district. During the 1998-1999 school year the African
American student population in each of these grades
consisted of 50 students in grade five, 42 students in
grade seven, and 32 students in grade 10. Of those, 75
students completed the SDQ (32 students in grade 5, 21
students in grade 7, and 22 students in grade 10).
Table 4.1 gives study participants by grade level.
Table 4.1
Study Participants bv Grade
Grade 5 Grade 7 Grade 10
N = 75 32 21 22
Results of the Study
The SPSS computer data analysis system was used to
analyze data. Hypothesis 1: Student' self-esteem scores
will differ according to grade levels. High school
students will have higher self-esteem scores than will
middle school and elementary students. Middle school
75


students will have higher self-esteem scores than will
elementary students.
in Table 4.2 students'self-esteem mean scores and
standard deviation, by grade level, are given. The
calculations were made using student responses to general
self questions on the SDQ*
Table 4.2
Self-esteem Scores as Measured bv the SDQ
NO. Students Mean Standard Deviation
Grade 10 N = 22 4.18 0.59
Grade 7 N = 21 4.44 0.52
Grade 5 N = 32 4.46 0.53
Results show that this hypothesis was not
supported. Self-esteem did not increase across grade
levels. Self-esteem scores of students in grade 5 were
higher than middle and high school students. Based on a
five point Likert scale self-esteem mean scores ranged
from 4.18 (grade 10) to 4.46 (grade 5). The results,
however, do support previous findings that African
American students maintain high self-esteem throughout
their childhood.
In order to interpret the importance of a measured
difference or relationship, it is now well accepted
76


that the educational or practical significance is just as
important as the statistical significance. Educational
significance is often expressed as an effect size, a
difference expressed in standard deviation units.
Effect sizes of the differences in self-esteem from
grade 7 to grade 10, from grade 5 to grade 7, and from
grade 5 to grade 10 are shown in Table 4.3. These effect
sizes represent the difference in mean scores divided by
the average (pooled) standard deviation of the three
groups. The pooled standard deviation is .55. According
to Glass, McGaw, and Smith (1981) "The most informative
and straightforward measure of experimental effect size
is the mean difference between experimental and control
groups divided by within group standard deviation" (p.
102). Glass et al. further state that magnitudes of
effect should not be associated with terms such as small,
moderate, or large. The "magnitudes of the effect will
gain meaning by reference to what is typical in similar
circumstances" (p. 104). Cohen (1962) on the other hand,
analyzed 70 studies to detect small, medium, and large
effect sizes and determined that the results produced
effect sizes of .20, .50, and .83, respectively (p. 151).
Table 4.3 gives the effect size comparisons for students
in grades 5, 7, and 10.
77


Table 4.3
Differences in Self-esteem
Expressed in Effect Sizes
Comparison Effect Size
Grade 10-Grade 7 i o
Grade 7-Grade 5 -0.05
Grade 10-Grade 5 -0.53
Results of effect size comparisons indicate that
not only was the hypothesis that high school students
will have higher self-esteem scores than elementary
students not supported but that the self-esteem of high
school students was substantially lower than the
self-esteem of elementary students. Within the context
of the effect size data, one might discern a
significantly higher self-esteem among both fifth grade
and seventh grade students than among tenth grade
students and only a very slightly higher self-esteem
among fifth grade students than seventh grade students.
Hypothesis 2: The relationship between
self-esteem scores and academic achievement among
African American students will differ according to
grade levels. Elementary students will have a
stronger relationship between self-esteem scores and
academic achievement than will middle school or high
78


school students. Middle school students will have a
stronger relationship between self-esteem scores and
academic achievement than will high school students.
In Table 4.4 students' self-esteem mean scores as
computed by responses to the SDQ's general self subscale
questions were correlated with students' total ITBS
scores and students' academic achievement (calculated
from students'semester report card grades in mathematics,
reading/language arts, science, and social studies). The
grade point averages in the four subjects mentioned were
computed to determine students' academic achievement.
Table 4.4
Correlations of Self-esteem
with Academic Achievement
Total ITBS Academic Achievement
Grade 10 -0.16 0.34
Grade 7 -0.23 -0.21
Grade 5 0.10 0.29
This hypothesis is not supported for students'
academic achievement (as measured by grades and total
ITBS scores). The correlation between academic
achievement (as measured by grades) and self-esteem in
79


high school show a trend in the direction expected but is
not significant. Results indicate that there is no
correlation between students' total ITBS scores and
self-esteem in grades 5, 7, or 10.
Hypothesis 3: The relationship between academic
self-concept scores and academic achievement among
African American students will differ according to grade
levels. High school students will have a stronger
relationship between their academic self-concept scores
and academic achievement than will middle and elementary
students.
In Table 4.5 students' achievement as determined by
semester report card grade averages in four subjects
(mathematics, reading/language arts, social studies, and
science) and students' Total ITBS are correlated with
students' responses to the SDQ's academic self-concept
subscale questions.
80


Table 4.5
Correlations of academic self-concept
with academic achievement
Achievement Total ITBS
Grade 10 0.37* -0.09
Grade 7 -0.17 -0.26
Grade 5 0.24 0.26
*2 < .1
This hypothesis was partially supported. Students in
grade 10 show a significant correlation between academic
self-concept and achievement (as measured by grades)
whereas no significant correlation between academic
self-concept and achievement was obtained in grade 7 or
grade 5. No correlation existed between students'
academic self-concept and students' total ITBS at any
grade level.
The relationship between academic self-concept and
achievement was explored in greater depth by examining
the correlation between academic self-concept in respect
to specific school subjects (reading and mathematics) and
actual achievement in these subjects. In Table 4.6
students' reading grades and students' ITBS reading
scores were correlated with students' responses to the
SDQ reading self-concept subscale scores.
81


Table 4.6
Reading Self-concept Scores Correlated with
Actual Readina/Lanauaae Arts Grades
and ITBS reading scores
Reading Grade/ Language Arts ITBS Reading Score
Grade 10 0.44* -0.12
Grade 7 -0.13 0.02
Grade 5 0.45* 0.11
*p < .1
Students in grades 5 and 10 show a positive
relationship between students' reading self-concept and
reading/language arts achievement (as measured by actual
reading/language arts grades). No correlation existed
for students' reading self-concept scores and students'
ITBS reading scores at any grade level.
In Table 4.7 students' mathematics grades and
mathematics ITBS scores were correlated with student
responses to the SDQ's mathematics self-concept subscale
questions.
82


Table 4.7
Mathematics Self-concept Scores
Correlated with Actual Mathematics Grades
and ITBS Mathematics Scores
Math Grade ITBS Math
Grade 10 0.01 0.25
Grade 7 Grade 5 0.13 1 o o to
No correlation existed between mathematics
self-concept and actual mathematics grades at any grade
level. No correlation existed between mathematics
self-concept and ITBS mathematics scores at any grade
level.
Summary
Self-esteem did not increase across grade levels as
hypothesized. Students in grade 5 showed higher
self-esteem scores than middle school and high school
students. Findings do support, however, previous
research that African American students report high
levels of self-esteem throughout their school years.
The relationship between self-esteem and academic
achievement as hypothesized was not supported. No
correlation was found between self-esteem and academic
83


achievement (as measured by Total ITBS and grade point
averages) in grades 5, 7 or 10.
The relationship between academic self-concept
scores and academic achievement as hypothesized was
partially supported. High school students did show a
significant correlation between academic self-concept and
academic achievement. Middle school and elementary
school students' results were not significant.
The relationship between reading self-concept
scores and academic achievement was explored. Results
indicate that significant correlations exited between
reading self-concept scores and actual reading grades at
the high school and elementary levels. No correlation
existed between reading self-concept scores and ITBS
reading scores any grade level.
No correlation existed between mathematics
self-concept scores and actual mathematics grades. In
addition, no correlation existed between mathematics
self-concept scores and ITBS mathematics scores.
The Interviews
Six students, two from each grade level, were
interviewed to gather in-depth information on how
students perceived their self-esteem, academic
84


self-concept, and academic achievement in relation to the
scores each received on the SDQ, ITBS, and grades for the
semester. The interviews explored the lack of a positive
relationship between academic achievement and
self-esteem. Students demonstrating a known discrepancy
between academic achievement and self-esteem were
interviewed. At each grade level, one student reporting
high self-esteem and receiving low grades and one student
reporting low self-esteem and receiving high grades were
interviewed.
Self-esteem mean scores and academic achievement
mean scores were used to determine which students to
interview. Three students (one from each grade level)
with self-esteem scores above the mean and academic
achievement below the mean and three students (one from
each grade level) with self-esteem scores below the mean
and academic achievement above the mean were interviewed.
Each interview lasted approximately 30 minutes.
Four of the interviews took place in the schools students
attended. The researcher was given private rooms, free
of distractions, to conduct the interviews. Two
interviews took place in the researcher's office. All
interviews were conducted, recorded, transcribed, and
coded by the researcher. An immediate rapport between
85


the researcher and each student appeared to exist. This
may be due to the fact that the researcher is also
African American. The researcher started each interview
by explaining the purpose of the interview and by
thanking each student for completing the SDQ. The
researcher stated that the interview would last
approximately 30 minutes and information students give
would be valuable in examining the relationship between
self-esteem and academic achievement and in gaining more
insight on what schools can do to increase African
American students' academic achievement. Questions asked
of each student are found in Appendix B.
All six students interviewed stated, without
hesitation, that they were satisfied with themselves.
Comments included "I'm satisfied with myself because I'm
me. I don't try to be like anyone else." One seventh
grade student summed up the general comments from all six
students by stating "Yes, I'm fine with myself."
Even though students with high self-esteem and low
achievement and students with low self-esteem and high
achievement were interviewed, similar answers to
self-esteem questions were given. Students responded
quickly and favorably when asked about their self-esteem.
86


This gives strong support to research findings that
African American students report high levels of
self-esteem. Also, the SDQ self-esteem responses show
mean scores for the entire population of students at each
grade level (using a Likert five-point scale) of 4.18 in
grade ten, 4.44 in grade seven, and 4.46 in grade 5 (see
Table 4.2).
When asked about satisfaction with their academic
achievement two of the six students reported that they
were not satisfied with their academic achievement. One
was achieving at high levels the other at low levels.
The student with low academic achievement reported that
grades could be better. Comments included: "I believe
that I could do better. I really didn't try in science
because I didn't like it. I really didn't try in a lot
of things." The student with high academic achievement
stated "In retrospect, I could have gotten straight As.
I think I've done well on standardized tests. I think
standardized tests are more about how well you can
perform under pressurehow well you can use
informationand how well you understand deductive
reasoning. I think that people who succeed are people
who work hard." The other students reported being
somewhat satisfied with their overall achievement. All
87


students reported that their academic achievement was
related to effort.
When asked to be more specific and respond to
whether their grades reflected their true academic
abilities all commented that they felt their grades could
be better if they tried harder. One high achieving
student stated that "in most cases grades reflect what
you really know. I think it is hard to get an A if you
don't know the information."
Students indicated that when they received an A it
was because they put forth the effort, focused, and did
the required work. Students indicated that the low
grades they received were because they felt they had not
tried hard, did not turn in their homework, and did not
complete the required assignments. These comments
indicate that students at each grade level felt that
grades reflect effort, not ability. One student's
comments sums up comments given at each grade level. "I
believe that effort is part of academic abilities.
Plenty of people are more intelligent than their grades
reflect. They don't care about the grade. I believe
that I should continue to try harder and put forth more
effort." Another student stated "I think people who
88


succeed cure people who work hard. Up until now, I have
worked fairly hard in school."
According to Marsh academic self-concept scores
show a strong, positive correlation to academic
achievement in specific content areas. Even though this
study did not obtain the high correlations that Marsh
found, students did report that they were not doing well
in specific subjects. Students were, however, satisfied
with their overall achievement. One student stated "I
get As in most of my subjects except math." Another
student stated "In language arts my grades are not that
good." Another student stated "Math, I've always kind of
struggled with. Language arts is more easy for me." When
discussing specific content areas students' perception of
how well they were doing in each subject generally agreed
with actual grades received. However, students reported
that in the subjects where they were receiving low grades
the grades were not an accurate reflection of their
ability but was due to their lack of effort.
The three students with high self-esteem and low
academic achievement indicated that their grades were
mainly average. They also indicated that it depended on
the subject. All three felt that if they tried harder
and stayed focused their grades would be better. The
89


three students with low self-esteem and high academic
achievement felt that their grades were good and stated
that they could be better. Students in tenth grade
indicated that they wish they had tried harder in middle
school.
Students believed they did "OK" on the ITBS. One
student achieving at high levels felt he had done well on
standardized tests because he was given opportunities to
participate in talented and gifted (TAG) classes in
middle school. He stated "When they put you in those
classes you are exposed to things that will be on
standardized tests. The first time I was introduced to
spatial reasoning was in seventh grade TAG classes."
This student further stated that "standardized tests are
more about how well you can perform under pressurehow
well you can use informationand how well you understand
deductive reasoning."
Students' self-esteem at each grade level appeared
to be high. Students related their satisfaction levels
to their participation in sports and their peer
relations. Students' overall academic self-concept
appeared to be high. Students reported accurately how
they were doing in specific academic subjects, but
attributed this to effort, and not to ability. They
90