Resilient adolescent females

MISSING IMAGE

Material Information

Title:
Resilient adolescent females investigating the relationships among gender role beliefs, coping skills, self-esteem, and academic achievement
Physical Description:
viii, 144 leaves : ill. ; 29 cm.
Language:
English
Creator:
Sands, Tovah
Publication Date:

Subjects

Subjects / Keywords:
Counselor Education thesis, Ph.D   ( lcsh )
Dissertations, Academic -- Counselor Education -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph.D.)--University of Florida, 1999.
Bibliography:
Includes bibliographical references (leaves 126-143).
Statement of Responsibility:
by Tovah Sands.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 030482061
oclc - 42458877
Classification:
lcc - LD1780 1999 .S221
System ID:
AA00017712:00001


This item is only available as the following downloads:


Full Text











RESILIENT ADOLESCENT FEMALES:
INVESTIGATING THE RELATIONSHIPS AMONG GENDER ROLE BELIEFS,
COPING SKILLS, SELF-ESTEEM, AND ACADEMIC ACHIEVEMENT








By

TOVAH SANDS


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1999




















This dissertation is dedicated with utmost love and
appreciation to my mother, Phyllis Durham,
and to my father, Norman Akiva Sands,
for their belief in me
over all these years.












ACKNOWLEDGMENTS

I have been blessed with tremendous support throughout the time I have spent

writing my dissertation. I am particularly grateful to my doctoral committee chair and

mentor, Dr. Mary Howard-Hamilton, for encouraging me to risk by going beyond what I

thought I could do. She encouraged me to write my first journal article, to teach my first

class, and to apply for the doctoral program. Her support and mentoring for the past

seven years have made all the difference to me.

My doctoral committee is simply the best. Dr. Martin Heesacker brought me into

the research community and role-modeled what excellent science can be about. He is an

inspiration second to none. Dr. Jim Archer gave me the opportunity to practice research

by inviting me to work with him in a partnership both educational and enjoyable. Finally,

Dr. Silvia Echervaria-Rafuls has been a steadfast ally. I am deeply grateful to all of them.

I wish to thank the many students, teachers, secretaries, school counselors, and principals

who participated in this research and who gave freely of their time to help complete this

study. Students and faculty from Eastside, Hawthorne, and Newberry High Schools in

Alachua County, Starke High School in Bradford County, Keystone Jr/Sr and Clay High

School Clay County, Interlochen High School in Putnam County, and P.K. Yonge

Developmental School of the University of Florida made this research possible.

It has been said that friends are the spice of life, but to me friends are a core part

of life itself. I could not have dedicated myself to my educational pursuits without the

many occasions of playing and laughter that my friends have brought me. From the










bottom of my heart I thank Agnes Ngoma and Michael Leslie, Lynne Goldman, Suni

Petersen, and Cynthia Moore.

Finally, it is my deep belief borne from personal experience that only by living

life in balance can one accomplish one's dreams. My father, Norman, mother, Phyllis,

step-father, Ray, and brother, Sheldon, have weathered the trials, tribulations, and joys of

my life with me; it is they who consistently urge me to balance my work with the hiking,

dancing, and socializing I enjoy so much. My family is my anchor and I am deeply

grateful to them. Last, but not least, I thank my son, Jason, for the patience he has

exhibited while I have been completing my education most of the years of his life. I could

not have done this without the many days and years of joy we have shared along the way.












TABLE OF CONTENTS


ACKNOWLEDGMENTS ................. ...................... iii

ABSTRACT ............... ................................ vii

CHAPTERS

I INTRODUCTION ................. ...................... .

Overview. ........................................... 1
Theoretical Foundations. ................................... 2
Statement of the Problem. .................................... 6
Need for the Study. ................................... ... 9
Purpose of the Study. ................................... 10
Significance of the Study. ................................. 11
Rationale...................................... 12
Hypothesis and Research Questions .............. ........... 14
Definition of Terms ................. ...................... 15
Organization of the Research ................................. 17

II REVIEW OF THE LITERATURE ........................... 18

Theoretical Model ....................................... 18
Females at Adolescence ................................... 23
Gender-Role Socialization ................................. 25
Gender Intensification ............................. ........ 26
Ethnicity and Gender ................... .................. 28
Resisting Stereotypes ...................................... 32
Academic Achievement ................. ................... 34
Decline in Academic Achievement ................. ..... .. 35
Academic Achievement and Gender-Role Socialization .......... 40
Middle School .............................. ......... 42
Differential Treatment in School ..... .............. .. 43
Self-Esteem ............................................. 45
R esilience .................... ................... ... .. 52

III METHODOLOGY. .................................... 60

Overview............ .. .................... ........ 60
Population .................. ................... .... 61
Sample and Sampling Procedures ............................. 61









D esign ........................................... ...... 63
Instrumentation ...... ................................... 64
Data Analysis .................................... ....... 72
Hypothesis and Research Questions ........................... 74
Methodological Limitations ................................ 76

IV RESULTS OF THE STUDY. ............................... 79

Descriptive Data .................. ................... .. 80
Hypothesis .................................. ....... 83
Research Question 1 ................. ..................... 84
Research Question 2 ................. ..................... 88
Research Question 3 .................. .................. .. 90
Research Question 4 ................. ..................... 92
Research Question 5 ................. .................... 93

V DISCUSSION. .......................................... 98

Hypothesis Summary and Explanation of Findings ................ 98
Research Questions and Explanations of Findings ................. 102
Theoretical Implications ................. ................... 110
Research Implications ................. ................... 1 I
Practical Implications and Recommendations ..................... 112
Limitations....... ............................. 115
Future Research .......................................... 115

APPENDICES

A INFORMED PARENTAL CONSENT FOR STUDY
PARTICIPATION ............. ................... 117

B PERSONALITY ATTRIBUTE QUESTIONNAIRE ............... 119

C ROSENBERG'S SELF-ESTEEM SCALE ........................ 123

D ADOLESCENT COPING ORIENTATION FOR PROBLEM
EXPERIENCES ............................... ....... 124

REFERENCES ................................................. 126

BIOGRAPHICAL SKETCH ............... ...................... 144














Abstract of Dissertation Presented to the Graduate School of the
University of Florida in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy

RESILIENT ADOLESCENT FEMALES: INVESTIGATING THE
RELATIONSHIPS AMONG GENDER ROLE SOCIALIZATION,
COPING SKILLS, SELF-ESTEEM AND
ACADEMIC ACHIEVEMENT

By

Tovah Sands

August, 1999

Chairperson: Mary Howard-Hamilton
Major Department: Counselor Education

The purpose of this study was to test a model of resiliency among middle

school adolescent females. Based on resiliency theory, the model predicted that

nonstereotypical gender role beliefs, high coping skills, and high self-esteem would

correlate with maintaining math and reading academic performance between fifth and

eighth grades. A second focus of the study was on examining the variables by race

and socioeconomic status.

A within-subject correlational design for survey research was employed. Path

analysis with manifest variables was used to test the hypothesis and two of the

research questions. The Personality Attribute Questionnaire, the Adolescent Coping

Orientation for Problem Experiences and the Rosenberg Self-esteem Scale were

administered and demographic and academic information was obtained for 291

participants.








Statistically significant results were found for some but not all of the research

questions and hypothesis. Self-esteem and math classroom grades correlated

significantly, as did the pairing of stereotypical gender role beliefs and healthy coping

skills with math classroom grades. Healthy coping skills and high self-esteem

correlated with each other, as did stereotypical gender role beliefs and healthy coping

skills. Math classroom grades had corresponding changes in math standardized test

scores, but reading classroom grades did not. Finally, variables were found to differ

significantly by race and socioeconomic status. African American young women

reported higher levels of self-esteem than did White participants. Math and reading

standardized test scores declined more significantly for African Americans than for

White students. Math and reading standardized test scores declined more significantly

for students from low-income families than from students from higher-income

families.

Overall, the findings offer mixed support for the proposed model of resiliency.

However, this research's results regarding young women's decline in math and

reading scores over middle school raises a red flag for school counselors, educators,

and parents to consider, and points to the need for further research in this area.













CHAPTER I
INTRODUCTION

Overview

The academic achievement of female adolescents declines among girls of every

ability level, race, ethnicity and socioeconomic level. National reports indicate that the

achievement levels of adolescent females in middle school, particularly on standardized

achievement measures, decline in comparison to their own performance on tests in

elementary school as well as in comparison to boys' test performance (American

Association of University Women [AAUW], 1992). This drop in test scores is

particularly significant because of the strong performance typically exhibited by girls in

the beginning of their educational career (Walker, Greenwood, Hart, & Carta, 1994).

When one considers the long-term implications for further educational attainment and

career development (Crockett & Crouter, 1995), it becomes apparent that depressed

academic achievement constitutes a distinct area of risk for adolescent females.

There is reason to believe that the intensification of gender messages that arrives

with adolescence and brings females such messages as, "Be pretty, not smart, and "Math

is hard," may be partially to blame for lowered academic achievement (Lindley &

Keithley, 1991; Orenstein, 1994). Although girls and women have made many gains in

educational attainment and career success in the past three decades, research indicates

that the influence of gender role socialization and resulting negative messages about

women and success continues to have a detrimental effect on girls' academic








achievement (Meece & Jones, 1996; Sadker & Sadker,1994). Young women today are

just as apt as previous generations to believe that doing well in school will make a girl

unpopular with boys (Suitor & Reavis, 1995). Girls learn that competing with boys--and

winning--can result in being left out of popular social circles and losing dates. Some girls

therefore conclude that it is better to play dumb in school than risk being unpopular by

appearing too smart (Harter, Waters, & Whitesell, 1997; Lips, 1993).

A second factor that may contribute to the fall of academic achievement in

adolescent girls is low self-esteem (American Association of University Women, 1992).

Adolescent girls have long demonstrated a disturbing drop in self-esteem that begins in

early adolescence and continues for many females throughout early adulthood (Gilligan,

1982; Gilligan, Rogers, & Tolman, 1991).

Resiliency theory may hold the key in describing the twin experiences of

declining academic achievement and declining self-esteem in adolescent females.

Although many young women find their academic achievement and self-esteem declining

in adolescence, a significant number of young women escape the hurdles of adolescence

unscathed and maintain their academic achievement levels (Gilligan et al, 1991; Harter et

al., 1997; Taylor, Gilligan, & Sullivan, 1995).

Theoretical Foundations

The construct of resilience has been defined as a protective mechanism that

modifies the way an individual responds to a risk situation and that tends to come into

action during critical turning points during one's life (Crockett & Crouter, 1995;

Garmezy, 1981; Rutter, 1987). Resilience can also be conceptualized as an individual's








ability to successfully cope with risk, stress, and adversity (Garmezy & Rutter, 1983;

Rutter, 1996).

The literature on resiliency has focused on investigating protective factors that

serve to buffer or protect an individual who is at risk from the negative results of the

particular risk factor to which they are exposed (Egeland, Carlson, & Srouffe, 1993;

Garmezy, 1983). For example, despite the difficulties associated with low socioeconomic

status, some young women are able to maintain both their self-esteem and their academic

achievement (Ward, 1996; Way, 1995).

In the past several decades researchers who have investigated the topic of

resiliency have focused on indicators of "at-risk" youth, the risks that put some young

people at great disadvantage (Masten & Coatsworth, 1998; Winfield, 1991). The second

focus has been on discovering and naming the protective factors that appear to mitigate

the risks for individuals who fit the at-risk criteria but somehow succeed in avoiding the

negative consequence of the named risk (Blocker & Copeland, 1994; Masten, Best, &

Garmezy, 1990).

The body of research on resiliency has grown beyond the original risk topics of

physical and mental illness, and poverty and racism, to include issues of educational

achievement (Winfield, 1991). In identifying protective factors educational researchers

have turned, for example, to describing African American adolescents who overcome

barriers of racism and disadvantaged schools to continue through high school, college

and professional career success (Copeland & Hess, 1995; Steele, 1998).

The notion that female adolescents face numerous challenges is hardly a new

concept. Many counselors, educators, and parents have worried that given the tremendous








pressures that adolescent females face today, and the landmines awaiting them if they

fall, being female and adolescent in today's culture is a more difficult venture than ever

before (Gilligan, 1982; Miller, 1976). Yet in the resiliency literature there is bare mention

of work investigating gender as a risk factor for young women, or of gender role

socialization as a correlate of risk for adolescent females.

The few studies that have included gender as a variable for risk, finding that it

indeed factors highly, can be found in literature discussing gender role socialization and

the development of young women (Gilligan, 1984; Grossman et al., 1992; Mirkin, 1994;

Taylor et al., 1995). Much of this research has been descriptive in nature. Leadbetter and

Way (1996), for example, describe interviews with urban adolescent girls of diverse

ethnic backgrounds who are self-confident and academically engaged.

This study utilized information, theories and research results from the literatures

of both resiliency theory and female adolescent socialization. This broad foundation

facilitated an investigation into the ways that protective factors of resilience seem to

buffer the stressors of gender inequity, enabling some young women to cope with the

stressors associated with gender role socialization.

Coping and stress theory stand on parallel tracks with resiliency literature. The

past few decades have seen a substantial increase in the literature describing the nature of

stress and coping in childhood and adolescence (Rutter, 1996). Empirical research has

documented the ways in which coping skills can reduce or alleviate the effects of stress in

adolescents (Copeland & Hess, 1995; Hauser et al., 1991). Researchers and clinicians

have evaluated the stress management strategies of at-risk youth in attempts to determine

the kinds of coping strategies that adolescents use that enable them to persevere in the








face of many different stressors. Effective coping skills can translate into enhanced

academic performance for young people (Spencer, Cole, DuPree, Glymph, & Pierre,

1993).

Much remains to be learned, however, in terms of the risk mechanisms that make

one person vulnerable to stress and not another, as well as individual differences in risk

exposure and the development of effective interventions (Rutter, 1996). This is the work

of resiliency studies such as this one.

In her theory on the development of women, Carol Gilligan (1982) states that

heightened pressure to conform to sex role socialization messages encourages adolescent

females to deny aspects of their own experiences, thoughts, feelings, and scholastic

abilities. As girls cross the threhold into adolescence, their self-confidence and self-

esteem decline (Brown & Gilligan, 1992; Orenstein, 1994).

An important construct found in the developmental literature that sheds light on

the experiences of young adolescents during the years of middle school is called the

gender intensification theory, the hypothesis that gender-differential socialization

accelerates during adolescence, especially for girls (Hill & Lynch, 1983). New domains

of behavior that were previously not impacted by gender rules for girls now become areas

in which gender-differential socialization pressure is exerted (Eccles & Bryan, 1994). A

critical byproduct of the crisis that many girls face at adolescence is their attitude and

behavior regarding schoolwork. In their effort to conform to stereotyped gender roles,

many girls suppress their academic ability and identity (Eccles, Barber & Jozefowicz,

1999; Wigfield & Eccles, 1994a).








Born from her work regarding women's need for connectedness, Gilligan (1982,

1984) describes a form of resistance in which girls and women go against the culture's

pressures and continue to establish and maintain interpersonal connections. In this theory,

resistance is the key to the development of a strong sense of self, i.e., the presence of

mutually respectful connection with others (Gilligan, 1982, 1984). In contrast to the loss

of self that is too frequently observed in young women, Gilligan describes healthy

resistance in which girls and women reject stereotypes that inhibit their achievements,

including those based on race, ethnicity, class, and sex (Fine, 1991; Taylor et al., 1995).

Theories of resilience and gender role socialization lay the theoretical foundation

for this investigation of adolescent females' academic achievement in middle school.

This research examined the relationships among the variables of self-esteem, gender role

socialization, and coping, and academic achievement in an effort to gain insight into the

ways that girls navigate through this critical period of life.

Statement of the Problem

Girls' academic achievement falls significantly during the middle school years,

especially in subject domains that are perceived as masculine (American Association of

University Women, 1992; Beaton, 1996; Hafner, 1990). The academic underachievement

of America's young women is a serious, widespread problem with long-term

ramifications that affect the future educational and career plans of girls and women, and

that carry associated ramifications for the productivity of the nation as a whole. Despite

significant strides within the last generation to overcome educational and career obstacles

for women and minorities, there continues to be discrepancy between what girls are

capable of achieving and demonstrate that they can achieve at early ages, and what they








actually achieve in high school, college, and in their careers (Crockett & Crouter, 1995;

Sadker & Sadker, 1994).

One of the consequences of females' lower test scores is the lowering of

expectations that girls have for themselves about their abilities, particularly in the areas of

math and science, with the result that many young women self-select themselves out of

advanced math and science courses in high school and college (Crockett & Crouter,

1995; Orenstein, 1994; Sadker & Sadker, 1994). College entrance itself is affected.

Lower test scores on traditional college admissions exams may deny the opportunity for

admission to certain colleges for women (Connor & Vargyas, 1992; Hyde, Fennema, &

Lamon, 1990; Mullis, 1991). Finally, girls and women often stay away from careers that

demand a high level of proficiency in math, science and technology, as noted by the low

number of women in the last part of this century who are employed in such occupations

as engineering and computer science (Hyde, Fennema, Ryan, Frost, & Hopp, 1990;

Sadker & Sadker, 1994).

As the new millennium begins, young women continue to exhibit declining

academic achievement, drop-out rates continue unabated, and too many young women

fall victim to serious problems ranging from truancy to substance abuse (Black &

Krishnakumar, 1998). Scattered efforts by some teachers in the area of gender equity in

schools and awareness by some counselors of societal pressures are important first steps

but are few and far between (McCracken, 1996; Meece & Jones, 1996). There is a

widespread lack of addressing the topic of sex role socialization as it impacts the

academic achievement of girls by school staff, counselors, parents, or society as a whole.








Until the root of the problem is examined, it is unlikely that these disturbing trends will

be reversed (Benokraitis, 1997).

When school counselors, mental health counselors, educators, and parents do not

understand nor respond appropriately to the factors that contribute to declining academic

achievement and self-esteem, girls' lives and futures are jeopardized (Orenstein, 1994).

For example, if 13-year-old Andrea believes that the boys that she is interested in at her

middle school think that it's really not cool for girls to do well in math, then it's highly

probable that Andrea will lower her standards for her own performance in math. And

when Andrea gets to high school, she may well opt out of advanced math courses as she

learns to comply with gender role socialization messages that increasingly come her way

(Kloosterman, 1990; Yong, 1992). If Andrea is lucky, she will have a significant adult in

her life, such as a parent, school counselor, mental health counselor or teacher, who is

aware of the tendency for girls' test scores to fall in middle school, who notices that

Andrea's standardized test scores fell in eighth grade, and who may even give her a pep

talk to "hang in there." However, chances are slim that Andrea's significant adult will say

to her, "Andrea, some boys you know may tease you because you're so good in math, but

I want you to know that there are some boys in your school who won't tease you, there

are some boys who think it's just as important for girls to do well in math as it is for boys

to do well in math. And most importantly, I want you to know that you can achieve and

succeed for yourself because you deserve to be all you can be" (Gilligan et al., 1991;

LaFrance, 1991).

This type of informed response will be much more effective for students like

Andrea than the standard comments she is more likely to encounter if she is fortunate








enough to have someone in her life who notices her academic decline, comments such as,

"What's wrong with you? You can do better than that." Adults who interact with young

women and who learn how to encourage girls to challenge negative gender stereotypical

influences and provide reliable role models may make all the difference in combating

declining school performance and falling self-esteem of adolescent females (Leadbeater

& Way, 1996; Way, 1998).

Need for the Study

The study of resilience in adolescent females brings hope for gaining valuable

insight into the challenges facing young women as they grapple with the strides and

limitations of women's greater entry into academia and the workforce. As the literature in

the past few decades has demonstrated, the study of resilience has revealed valuable

insights into a number of high risk areas including racism, poverty, and illness

(Montgomery et al., 1993; Pungello, Kupersmidt, Burchinal, & Patterson, 1996; Richters

& Martinez, 1993). Therefore, an important next step in this work is to apply the

principles of what has been discovered about resilience to the experience of adolescent

girls during an often tumultuous and critical period of their lives.

Differences in the experiences of adolescent girls need to be investigated by race

and class. There is little empirical research investigating class and race-related variations

of stress and coping during adolescence (Bush & Simmons, 1987; Spencer & Dornbusch,

1990), or on gender role socialization (Hagen, Paul, Gibb, & Wolters, 1990). For

example, the UUAW (1992) study reported distinct ethnic differences in the levels of

female adolescents' self-esteem. In contrast to White and Hispanic young women,








African American females retain high self-esteem that actually increases at the

completion of high school (Erkut, Fields, Sing, & Marx, 1996; Fordham, 1993).

Stories of resilient youth have been described in the qualitative literature

(Leadbeater & Way, 1996; Taylor et al., 1995; Way, 1998). However, there is little

empirically based research regarding resilient adolescent females. Quantitative studies

have an important role in the study of resilience by investigating larger numbers of

adolescents to help establish the likelihood that certain traits associated with resilience do

indeed contribute positively to the risk condition.

Although there is general consensus that coping skills are important for young

people, there has been little research examining the relationship of coping skills to gender

role socialization (Nolen-Hoeksema, 1995). The possibility that young women with good

coping skills may be better able to resist disparaging notions of female academic pursuit

and achievement should be investigated. Assisting girls to develop successful coping

strategies to resist negative stereotypes that discourage them from doing their best in

academic pursuits could be key for counselors who work with adolescent females.

Purpose of the Study

The purpose of this study was twofold: to identify risk factors that impede

adolescent females from maintaining academic achievement, and specifically to

investigate whether or not gender role socialization messages could be identified as a risk

factor for adolescent girls; and secondly, to identify protective factors that buffer

adolescent females from the risk of declining academic achievement over middle school.

This was accomplished by examining the relationships among adolescent females'

academic achievement and traits of resilience in young women, including high








self-esteem, healthy coping skills, and girls' ability to resist negative stereotypes based on

gender. In order to gain as realistic and applicable an understanding of this topic as

possible, this study examined each research variable by race and socioeconomic status.

In this study I proposed that resistance to gender role stereotyping be identified as

a protective factor that serves to help adolescent females maintain their academic

achievement. The study's research questions investigated corollaries of the primary

hypothesis by posing questions such as, How do girls of different races and

socioeconomic backgrounds differ in making use of protective factors? Do girls who

resist gender role socialization pressures have other characteristics of resilience in

common?

There is a critical need for mental health counselors and school counselors to

develop and implement effective intervention strategies to address the continuing decline

of academic achievement by middle school girls. By understanding the role that gender

role socialization plays in the academic achievement of adolescent girls and by utilizing

strategies gained from the study of resilience in adolescent youth, counselors will be

better able to address and assist young women in reversing the decline in academic

achievement that is demonstrated by too many adolescent females.

Significance of Study

Decreased academic achievement for girls and women has been linked to

diminished career opportunities and restricted quality of life (AAUW, 1992; Sadker &

Sadker, 1994). Many observers agree that gender related factors are involved with the

decrease in academic achievement by adolescent girls (Reis, 1991; Rigsby, Stull, &








Morse-Kelley, 1997), yet the fundamental messages that contribute to this phenomenon

have been either ignored or relegated to back-seat concerns.

The period of middle school is a time of high risk, and the transitions from grade

school to middle school, and from middle to high school are widely recognized as

difficult periods for adolescents (Winfield, 1995). What characteristics do girls who are

able to maintain their academic achievement during middle school have in common?

What can be learned from these resilient girls?

Rationale

This research is unique in its focus on stereotypical gender role beliefs as a risk

factor for adolescent females as conceptualized within resilience theory. Although

previous studies that have identified stereotypical gender role socialization as a risk, they

have done so in literature outside of the main body of resiliency literature. Another

unique contribution of this study is the analysis of findings by race and socioeconomic

status; few research studies have examined these issues by these critical demographic

components.

This research sought to directly address gender role socialization as a factor that

puts a young woman at risk from the problem of declining academic achievement in

middle school. In this study, low femininity on the Personal Attribute Questionnaire

(PAQ), the frequent use of healthy coping skills, and high self-esteem were utilized as a

way of documenting resilience in youth. The PAQ, developed by Spencer and Helmreich

(1978), is a widely respected assessment tool that measures sex role stereotypes by

assessing the respondents' beliefs that male and females differ in many of their

characteristics.








The use of healthy coping skills were assessed by the Adolescent Coping

Orientation for Problem-Experiences (A-COPE), developed by Patterson and McCubbin

(1987) specifically for adolescents ages 13 to 19 years of age. The A-COPE provides a

measure of self-reported coping strategies such as venting feelings, crying, or seeking

spiritual support. An advantage that the A-COPE brings to this study is that it is based on

the theoretical conceptualization of an integration of individual coping theory and family

stress theory. A-COPE has been used effectively in a number of studies that seek to

assess resilience in at-risk adolescent youth (Copeland & Hess, 1995; Plancherel &

Bolognini, 1995).

Self-esteem was assessed via Rosenberg's Self-Esteem scale (1965). This scale

has been used extensively to assess adolescents and was originally developed for use with

high school students. Reports indicate high reliability and validity. It has been used

frequently in previous research on resilience in youth (Goodrich, Rampage, Ellman, &

Halstead, 1988; Grossman et al., 1992).

An important aspect of this research is the attempt to consolidate relevant

previous research and writings in various areas that may be of help in understanding and

assisting in the academic achievement of adolescent females. Previous research findings

in the areas of gender role socialization and resilience have important implications that

need to be considered when looking at the declining self-esteem and academic

achievement of adolescent girls in middle school. Previous research studies have linked

self-esteem with academic achievement among girls, and other researchers have found

that good coping skills are related to higher academic success among adolescents. Still

other studies have looked at gender role attitudes and academic success. In this study I








examined each of the variables and their relationships with each other through theory and

empirical research. Data were analyzed and variables correlated so that the strength of the

relationship among the independent variables--self-esteem, gender role socialization, and

coping--were correlated with the dependent variable of academic achievement, and then

examined by race and socioeconomic class.

Much of the research to date on resilient youth has either been theory-

development work or qualitative studies of small populations (Copeland & Hess, 1995).

This study will add to the resilience literature by employing a quantitative research

design utilizing a fairly large sample of African American and White female adolescents

of differing socioeconomic backgrounds.

Hypothesis and Research Questions

Hypothesis

As stereotypical gender role beliefs among early adolescent females increase, and

as healthy coping strategies and self-esteem decrease, it is predicted that academic

achievement scores from fifth to eighth grade will decrease.

Research Question 1

How do the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem correlate with each other?

Research Question 2

Which pairs of the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem have the greatest impact on the criterion variables of fifth and

eighth grade test score differences and fifth and eighth grade classroom grade

differences?








Research Question 3

What is the relationship between the two criterion variables of fifth and eighth

grade standardized test score differences and fifth and eighth grade classroom grade

differences?

Research Question 4

Will the relationships among stereotypical gender role beliefs, coping skills, self-

esteem, and academic achievement for middle school females vary by race or

socioeconomic status?

Research Question 5

Will the mean scores of stereotypical gender role beliefs, coping skills, self-

esteem, and academic achievement for middle school females vary by race or

socioeconomic status?

Definition of Terms

For the purpose of this study, the following definitions are presented:

Academic achievement is operationalized for this study as performance on

standardized tests and as classroom grades.

Coping is one of many resources that can add to or decrease the amount of stress

that adolescents experience. Healthy adolescent coping is viewed as achieving a

successful balance between meeting demands of the self, the family, and the community

(Patterson & McCubbin, 1987).

Expressiveness is defined as comprising interpersonally oriented (e.g., gentle,

sympathetic, understanding) personality traits (originally termed femininity) (Spence,

1984).








Femininity is defined as comprising interpersonally oriented (e.g., gentle,

sympathetic, understanding) personality traits (also termed expressiveness) (Spence,

1984).

Gender intensification refers to an acceleration of gender-differential

socialization during adolescence, typically occurring at the onset of puberty that is

particularly acute for girls (Hill & Lynch, 1983).

Gender role messages are the unwritten rules or messages regarding gender roles

that men and women use to guide their lives (Worell & Remer, 1992).

Gender role socialization refers to generalized rules of living that females and

males learn by observing the differential treatment and responses that society relegates to

each sex (Worell & Remer, 1992).

Instrumentality is defined as comprising socially desirable self-assertive (e.g.,

willing to take risks, dominant, self-reliant) personality traits (originally termed

masculinity) (Spence, 1984).

Masculinity is defined as comprising socially desirable self-assertive (e.g.,

willing to take risks, dominant, self-reliant) personality traits (also termed

instrumentality) (Spence, 1984).

Racial Socialization refers to a proactive orientation toward racial barriers that

African American families teach their children, including ethnic pride and self-

development (Ward, 1996).

Resilience is defined as a protective mechanism that modifies the individual's

response to a risk situation and operates at critical turning points during one's life,

i.e., the ability to cope successfully (Rutter, 1987).








Self-esteem is the individual's overall evaluation of herself. Also termed global

self-esteem, it includes whether or not a person is generally satisfied with his or her life,

considers themselves worthy, and holds a positive attitude toward their self (Rosenberg,

1979).

Sex roles/Gender roles are used interchangeable in this study to indicate societal

expectations concerning the appropriate behavior for men and women (Worell & Remer,

1992).

Stress is "a particular relationship between the person and the environment that is

appraised by the person as taxing or exceeding his or her own resources and endangering

his or her well being" (Lazarus & Folkman, 1984, p. 19).

Organization of the Research

This study is reported in five chapters. In Chapter II, the related professional

literature is reviewed. This review includes (a) a discussion of the theoretical model of

resilience, (b) an overview of early adolescence for young women, and the relationship of

ethnicity, socioeconomic level and gender role socialization to the female experience of

adolescence, (c) academic achievement, the middle school context, and differential

treatment in school, (d) self-esteem, and (e) resiliency literature. Chapter III describes

the research methodology, data collection and analysis methods. The data analysis and

results of the study will be presented in Chapter IV. A summary of the study, discussion

of the results, conclusions drawn from the research, limitations of the study,

recommendations for future investigations, and suggestions for practical application will

be provided in Chapter V.













CHAPTER II
REVIEW OF THE LITERATURE

Theoretical Model

This study was based on resiliency theory as applied to adolescents. In this section

the theoretical basis of resilience will be discussed. The section will conclude with an

explanation of the research proposal to adapt resiliency theory to include gender role

stereotypes as a risk, and resistance to gender role stereotypes as a protective factor.

The roots of resilience theory can be traced back to the construct of risk, which

has a long history in the medical and psychiatric literatures. In areas as diverse as

schizophrenia and response to trauma and disaster, the term risk was used to specify

conditions that make individuals susceptible to disease or mental disorders (Garmezy,

1983, 1987; Masten, Best, & Garmezy, 1990; Rutter, 1979). The medical literature

chronicles examples of adaptive functioning in individuals who were exposed to trauma,

and it was from this foundation that theories of resilience were proposed (Cicchetti &

Garmezy, 1993).

For the past two decades research has focused away from illness and pathology

toward a focus on resilience, conceived as individual variations in the ways that people

respond to risk, stress, and adversity (Kobasa, 1979). In other words, some individuals

are resilient and cope successfully while others do not (Rutter, 1987).

The roots of resilience in social science literature is often credited to the work of

Garmezy and his colleagues who began investigations of individuals who are competent








in the midst of stress (Garmezy, 1981). Two decades ago Garmezy first used the term

"invulnerability" to characterize a group of Black children who, despite exposure to

significant stress associated with ghetto poverty and prejudice, had become competent

and well adjusted.

These writings were among the earliest examples of efforts to emphasize the

importance of examining protective factors in "at-risk" populations. Protective factors are

conceptualized as characteristics in an individual's world that mitigate against the

development of behavioral and psychological problems despite the existence of risk

factors (Rutter, 1979). A major focus of the resiliency literature has been to identify

protective factors in children's lives.

The resiliency literature has borrowed heavily from theories on stress, coping, and

hardiness. Garmezy (1983) and Rutter (1983) worked from a stress and coping

perspective to develop the idea of protective factors. Resilience has also been

conceptualized as an integration and application of individual coping theory and family

stress theory (McCubbin & McCubbin, 1996). From this perspective, the adolescent is

viewed as needing to simultaneously manage demands that are individual, family and

community based; successful coping is defined as simultaneously being able to mange

demands of one's self and one's family and the community of peers, school, and society

(Patterson & McCubbin, 1987).

There is a fine but distinct difference between the theoretical underpinnings of the

constructs of hardiness and of resilience. Hardiness can be conceived of as an identifiable

set of personality characteristics possessed by an individual that operates at all times in

the life of the individual (Kobasa, Maddi, & Kahn, 1982). In contrast, resilience is the








notion of a set of characteristics, some of which are similar or even identical to hardiness

traits, that operate in the face of risk. In the conceptualization of resilience, the presence

of risk is fundamental and is the hedgepin that sets the process of resiliency in motion.

Resilience emphasizes the recovery process, and has been described as the ability to

recover one's strength or spirit (Gordon, 1997).

In addition, research into childhood resilience is making increasingly clear that

resilience is not an innate characteristic, but depends upon a combination of factors

including attributes of the child, socialization experiences both within and outside the

family, and interactions between these components (Cowen, Wyman, Work & Parker,

1990; Rutter, 1990). Intervention strategies that arise from the context of resilience

literature urge that strategies be implemented to both decrease the child's exposure to the

stressor(s) and to increase the number of protective factors (Gordon, 1997).

Several concepts were borrowed from the hardiness literature for the development

of resilience theory. Kobasa, Maddi, and Courington (1981) and Kobasa, Maddi, and

Kahn (1982) proposed that the three hardiness factors of control, commitment, and

challenge decreased the harmful effects of stressful life events. These factors were

described in the context of resilience as (a) a hardy or resilient person's attempt to

influence the course of an event (control), (b) curiosity about how the event happened

and dedication to what it is (commitment), and (c) trying to learn from the event in a way

that will enhance personal growth (challenge) (Blocker & Copeland, 1994).

There is some disagreement in the resilience literature as to just what are risk or

protective factors, and widely varying definitions of each are used across studies

(Egeland, Carlson, & Srouffe, 1993). Although resilience is often conceptualized as an








individual construct, protective factors may include variables from the individual, family,

and social environment. Some of the individual factors described in studies as providing

protection to adolescents from risk include (a) self-esteem (Garmezy, 1983; Rutter,

1979), (b) curiosity (Garmezy & Rutter, 1983), (c) internal locus of control (Garmezy,

1987), (d) intelligence, (e) coping style, (f) social skills (Garmezy,k 1983; Rutter, 1983),

and (g) a sense of self (Masten & Coatsworth, 1998). Sense of self has been described as

a sense that one is able to exercise a degree of control over the environment and one's

life, a trait shared with hardiness descriptors (Blocker & Copeland, 1994).

Context is an important aspect of conceptualizing resilience. Contexts important

to adolescents such as the family, school, peer group, and local neighborhood play a key

role in the opportunity or lack thereof for risk and protective factors to develop.

Adolescents live in multiple contexts that shift at times, and these shifts, or transitions,

often represent periods of increased stress or risk for the adolescent. The transition into

middle school is one example of such a contextual transition period (Bronfenbrenner,

1989).

The educational research on resilience is not uniformly defined. The term

"at-risk" became popularized in education after the publication of A Nation at Risk

(National Commission on Excellence in Education, 1983), when educators, policy

makers, and researchers applied the term fairly indiscriminately to refer to any youth who

were likely to experience school failure, teen pregnancy, or some other negative

developmental outcome. Often educational definitions of at-risk utilize demographic

characteristics such as racial and ethnic background and social class as indicators. This








becomes a short-hand, if inaccurate, way for schools and agencies to identify students

who may be in need of additional academic or social services (Winfield, 1991).

Critics point out that using the label of at-risk based on demographic

characteristics serves to shift attention away from the social conditions that place

adolescents at risk, and tends to locate the risk within the adolescents themselves

(Steele, 1998). By shifting the burden of change to the adolescent it thus relieves the

larger society of responsibility for addressing the inequities of race, class, and gender that

create the conditions of risk (Fine, 1991; Taylor et al., 1995).

The critical issues in resilience in the educational realm center around identifying

the protective processes and mechanisms that reduce risk of educational failure and foster

resilience in academic achievement and career success. Questions have been posed

regarding how protective processes operate at different developmental levels or transition

points in the schooling process (Crockett & Crouter, 1995; Winfield, 1995). It is

important to understand whether or not resilience processes are the same for different

ethnic and socioeconomic groups. Researchers are anxious to bring concrete suggestions

to schools, administrators, teaching, community groups, and policymakers regarding

what can be done to enhance and foster the development of resilience in young people

(Winfield, 1991).

In this study careful reference was made to the use of the at-risk label. In this

research I was not proposing that female adolescence per se be considered a risk factor,

although some authors have suggested this. Rather, in this research I proposed that the

factor of high femininity, or subscribing to gender stereotypes by female adolescents, be

considered a risk factor that predisposes young women to detrimental consequences,








including lowered academic achievement. The proposed protective factor of resistance to

stereotypical gender role beliefs was investigated concurrently.

The interaction of high femininity with self-esteem and coping skills may also

influence academic achievement for young women. It is for this reason that self-esteem

and coping skills were operationalized for this study. Researchers urge utilizing more

than one behavioral index to assess resilience because although an adolescent may appear

to be adapting positively within the school arena, for example, the same young woman

may display poor adaptation in another context (Luthar, Doernberger, & Zigler, 1993).

Females at Adolescence

This section includes a summary of studies that describe the period of

adolescence, gender-role socializatio,n and gender-intensification, with particular

attention paid to the ways in which these factors may become risk factors for female

adolescents. The ways in which the variables of this research impact female adolescents

of color are discussed next, concluding with studies that focus on young women who

resist gender-defined stereotypes.

Adolescence is a time of tremendous change sometimes referred to as a double-

edged sword (Lerner, 1993). On the one hand, adolescence is a time of burgeoning

possibilities with new capacities for complex thinking, for self-understanding, for

learning about one's sexuality, for more mature relationships with parents, peers, and

teachers, and for beginning to consider the world of work and careers (Crockett &

Crouter, 1995; Eccles & Bryan, 1994).

On the other hand, adolescence can also a time of profound biological,

psychological, and social risk (Natriello, McDill, & Pallas, 1990; Plancherel &








Bolognini, 1995). Some observers go so far as to warn that half of America's adolescents

are at moderate or great risk, defined as engaging in unsafe sexual behaviors, teenage

pregnancy, teenage childbearing, substance abuse, school failure and dropout, juvenile

delinquency, and school underachievement (Dryfoos, 1990; Eccles & Bryan, 1994).

When adolescents face multiple changes simultaneously (e.g., when menarche occurs at

the same time as a school transition), the risk is even greater that problems will occur in

the youth's development (Lerer, 1993).

During adolescence, young women receive powerful gender-role socialization

messages from adults in their lives, from peers, and from the culture at large that exert

tremendous pressure for them to conform to stereotypical ideals of femininity (Brown &

Gilligan, 1992; Eccles & Bryan, 1994; Gilligan et al., 1991). According to Gilligan

(1991), one of the most common tactics that adolescent females employ in order to

conform to gender role expectations is the strategy of self-silencing. By suppressing their

involvement in the classroom, girls believe that their peers will accept them.

An illustration of self-silencing can be found in the following study. Using an

intensive, interview-based, longitudinal design, Gilligan and her colleagues (Gilligan et

al., 1990) studied a group of nearly 100 girls between the ages of 7 and 18 years of age as

they moved from middle childhood into and through adolescence. The results indicated

that until the age of 12 girls possess a wealth of self-confidence and courage and seem

resistant to harmful notions of feminine behavior. However, as the girls in the study

matured, they became more hesitant and uncertain. They spoke of a tension between their

own experiences of the world and societal messages they receive regarding gender-role

expectations.








Jean Baker Miller (1984) described the experience of adolescence for females as a

period when young women receive very strong messages from society that they should

not fully utilize all of their capacities. They learn that females should always put others'

needs or desires before their own. The AAUW (1992) report provides some evidence for

the conflict that young women face in their report of a 13-point drop in girls' self-esteem

between the ages of 9 and 15.

Gender-Role Socialization

Socialization in American culture is a different process for girls than for boys.

Girls are taught by society to focus on personal relationships and to develop sensitivity,

empathy, warmth, and concern about others. Young women are encouraged to

overemphasize physical appearance, dating, popularity, intimacy, emotionality, and other

behaviors that are directly associated with the female role of wife and mother (Chodorow,

1978; Suitor & Reavis, 1995).

Today's generation of adolescent females may be less likely than their mothers to

hear messages directly stating women's inferior academic abilities, but they are more apt

to receive messages of a subtle form of sexism (Benokraitis, 1997). For example, today's

young women may see fewer images on television of women staying at home and raising

children than their mothers did, but they also will see few, if any, images of women in

traditionally male occupations such as a chemistry or physics (Lips, 1993). And the

images they see of high-status professional women will frequently be coupled with strong

sexual overtones, such as the sexy, young female lawyer portrayed on the popular

television program, "Alley McBeal."








Debate lingers regarding what causes socio-behavioral gender differences.

Essentialists argue that the differences in developmental patterns between girl and boys

are due to biology. The constructivists, on the other hand, emphasize the influence of the

environment and the social institutions that shape development and are exemplified in the

work of Gilligan (1982) and Kegan (1982). Bem (1985) theorized that children develop

gender schemas to help them guide their behavior in the world. Gender stereotypes arise,

some speculate, through observation of the world (e.g., secretaries are usually female),

and through explicit teaching, leading both sexes to believe that socialized behaviors are

natural for each sex (Halpern, 1997).

Research studies have borne out the prevalence of gender-role stereotypes that

persist at the turn of the 21st century. In a longitudinal study of low socioeconomic class

African American youth from four middle schools in a large southeastern urban city,

researchers examined gender-linked beliefs and attitudes. Data were gathered from 394

male and female youth utilizing the Personal Attributes Questionnaire and the Iowa Tests

of Basic Skills. Results indicated the presence of many stereotypical gender-linked

beliefs and attitudes including the belief by both males and females that the

characteristics necessary for successful educational attainment are masculine. This study

seems to indicate that young women recognize the limitations of gender stereotypes but

they do not want to waver too far from them, even at the expense of their own declining

academic achievement (Swanson & Spencer, 1997).

Gender Intensification

The gender intensification theory, first proposed by Lynch (1983), states that








gender-differential socialization accelerates during adolescence. Typically appearing at

the onset of puberty, this intensification is particularly acute for females. Gender

intensification is blamed for the peer pressure girls feel; the peer group is the principle

enforcer of gender-rules for youth. Gender intensification is thought to lead to heightened

self-consciousness and a decrease in self-esteem (Hill & Lynch, 1983).

One of the consequences of gender intensification is that certain activities become

gender labeled that previously were not, and adolescents experience increased pressure to

conform to this labeling (Eccles & Bryan, 1994; Hill & Lynch, 1983). As gender-role

appropriateness increases in importance for early adolescents, young women tend to have

less positive beliefs about their capabilities and to be less involved in activities that they

see as inappropriate for their gender (Eccles, Barber & Jozefowicz, 1999; Wigfield &

Eccles, 1994b).

According to the tenants of the gender intensification theory, the display of

competence for females becomes antithetical in the gender labeled world of 'his' and

'hers'. Girls are no longer highly rewarded for their achievements, and academic success

is often perceived as incompatible with more highly regarded traits such as popularity

and sociability (Hill & Lynch, 1983).

Research studies indicate that girls tend to develop a more gender-role

stereotyped view of academic subjects as they move through secondary school (Eccles, et

al., 1989; Handley & Morse, 1984). In a study of girls in 10th, 11th and 12th grades, Eccles

(1994) found that as girls reached their senior year of high school, they endorsed more

stereotypical views of math and English than did the younger girls (Eccles & Bryan,

1994). The results of this study seem to indicate that young women incorporate gender








stereotypes about their lack of competence in particular subjects despite evidence to the

contrary (Eccles & Bryan, 1994).

In an early study investigating adolescents' tendency to gender-label academic

pursuits, researchers (Stein, 1971; Stein, Pohly, & Mueller, 1973) asked students in 2nd,

6th, and 12th grades to rate six achievement areas as more feminine or more masculine.

The students rated the areas from feminine to masculine in the following order: social,

artistic, reading, arithmetic, spatial, and mechanical. In a follow-up study of sixth and

ninth graders, Stein (1971) reported that gender labeling was much stronger among ninth

than sixth grade students, with girls placing more importance on feminine areas and boys

on masculine areas.

Ethnicity and Gender

The contribution that ethnicity makes in the lives of young women of color is

crucial for understanding the nature of risk and resilience. For White girls, gender may

indeed be key in their struggle for personal identity and social place. For girls of color

and girls living in poverty, however, gender is likely not the only, nor necessarily the

most salient, struggle in which these adolescent females engage (Erkut, Fields, Sing, &

Marx, 1996; Phinney, 1989).

Likewise, the importance of context--the broader societal conditions and events

that impinge upon families--must be considered in order to gain an understanding into the

motivations, successes, and life courses of adolescents. Ecological factors such as cultural

stereotypes, school experiences, and peer relations figure prominently in a young

woman's attainment of academic success (Pungello, 1996). Spencer (1995) theorizes that

self-esteem is linked to the experiences of risk associated with societal stereotypes. These








risks impact experiences of stress, coping methods employed, and outcomes such as

academic success (Steele, 1998).

Despite the need for studies that reflect the full diversity of cultures and

experiences present in the United States, most of the social science literature written

about adolescent resiliency has focused on African-American and White adolescents.

Much of the discussion in this section will therefore focus primarily on these populations

of adolescent females.

An important facet in examining stereotypical gender-role adoption by diverse

populations of adolescent females is the question of whether or not the feminine

stereotype is actually a stereotype of middle-class White females that does not apply to

women outside of this population. In an investigation of this question, college students

were asked to state a stereotype for each of four groups of racially diverse women. The

authors found that although the stereotypes differed significantly by both race and social

class, with White women and middle-class women being described in ways most similar

to the dominant culture's traditional stereotypes of women, all four groups were rated in

ways that are consistent with the dominant culture's stereotype of femininity. This study

appears to indicate that young women of all ethnic groups are subject, with variations

peculiar to that group, to stereotyped gender socialization messages (Landrine, 1985).

Other studies examining gender stereotyping and gender-role attitudes by racial

group show mixed results. Reviews of research regarding gender role attitudes among

African Americans finds some studies in which Black females and males are more

egalitarian than Whites in their gender-role attitudes and some in which the opposite is

found (Hatchett & Quick, 1983). For example, in a comparison of the gender-role








attitudes of American Black and White women, Dugger (1991) found that Black women

were more likely than White women to reject stereotypical views of women's roles.

Black women were more likely than White women to observe sex discrimination, support

the women's' movement, admire intelligent, outspoken women, and accept nontraditional

family structures. On the other hand, these Black women were more likely than White

women to accept that girls and boys should be brought up to be feminine and masculine,

respectively (Dugger, 1991).

Some authors who consider the topic of the academic achievement of African

American female adolescents warn that one must recognize that in a culture that

continues to adhere to institutional racism, young women of color must contend with a

greater number of variables in order to attain academic achievement than do White

female students (Fordham & Ogbu, 1986). Not only are attributes such as intellectual

ability, aspirations, achievement motivations, and personal and social identity important

for achievement, but the social environment of the school and the adolescent's support

network is also critical. The school as a social institution reflects and transmits

mainstream culture. In a society with vestiges of racism, some scholars argue that the

impact of the school environment can make the difference between failure and success

for adolescents of color (Fordham, 1993; Steele, 1998; Steinberg, Dornbusch, & Brown,

1992).

Research in the area of adolescent achievement frequently reports findings of

ethnic differences in school performance. Many studies over the past few decades have

reported that African American students generally earn lower grades, drop out more

often, and attain less education than do Whites (Steinberg et al., 1992), although there is








some indication that differences on standardized achievement tests between African

American and White school children are slowly decreasing (Montgomery et al., 1993).

There is little consensus about the causes of test score differences between

African American and White children. Some studies indicate that higher levels of poverty

among African Americans may account for some of the achievement differences, but

other studies find the opposite results (Ogbu, 1981). Additional factors that may effect

academic achievement include variations in school district expenditures, with more

affluent districts tending to have greater resources including more technology and more

advanced-level courses (Rigsby, Stull, & Morse-Kelley, 1997). Pungello et al. (1996)

examined the long-term effects of low family income and stressful life events on math

and reading achievement for 1,253 low income African American children across Grades

2 through 7. Risk effects were found for math achievement but not for reading

achievement. This research points out the many risk factors that African Americans often

face, including the effects of racism even in the absence of poverty or other stressful life

events (Pungello et al., 1996).

Peer groups are very influential for Black students' success. In some cases peer

groups act as a stressor, exerting a negative influence on academic learning. Peer groups

may believe that teachers and curriculum represent White culture, thereby influencing

young women to forfeit their studies or else be seen by peers as cooperating with the

enemy. Fordham and Obgu (1986) reported that the Black adolescents they interviewed

defined studying and doing well academically as "acting White," and that this notion

affected competent African American young women by reducing their motivation to

make the necessary effort to perform well in school.








Resisting Stereotypes

"Some girls develop under the most adverse conditions, but the interesting

question to me is, Under what conditions do most girls develop to their fullest?" (Pipher,

1994, p. 293).

Recent studies are providing documentation of adolescents who exemplify

resilience, living in adverse circumstances and yet demonstrating their own version of

protective factors. In a longitudinal qualitative study with 12 girls and 12 boys from an

urban, ethnically diverse high school, Way (1998) chronicled interviews she conducted

over a three year period. She recorded the assertive voices of adolescent girls who could

be called "political resisters," although they did not fit the stereotype of loud, bold,

African American inner-city girls (Fordham, 1993). These young women were able to

express their anger and frustration, feelings often unspoken by less resilient females. The

courage that it takes for these adolescent females to forcefully express themselves may be

considered demonstration of a protective factor.

Young women's ability and willingness to speak up for themselves can be viewed

as an indicator of self-esteem and as a protective factor. The research findings of Harter,

Waters, and Whitesell (1997) indicate the tremendous variations in what Gilligan (1982)

termed "level of voice." In cross-sectional research examining approximately 600 middle

and high school primarily White students of varied socioeconomic status, the researchers

reported that many young women self-reported that they do engaging in expressing their

opinions. Their research did not find lack of voice among female adolescents as a group,

but instead, found marked individual differences in level of voice among both male and

female adolescents. The authors examined the hypothesis that expression of voice will








differ according to adoption of a feminine orientation, as measured in part by the

Personal Attributes Questionnaire and the Bem Sex Role Inventory. Their research found

that approximately 25% to 35% of their sample scored high in femininity, while

approximately 60% to 70% scored as androgynous. Feminine girls reported significantly

lower levels of voice than did androgynous girls in the public context of schools,

particularly with teachers and classmates (Harter et al., 1997).

Recalling that the convergence of multiple transitions is particularly stressful for

adolescents, some studies indicate that gender stereotyping may abate when the intensity

of multiple changes and stressors decline. Alfieri, Ruble, and Higgins (1996) investigated

the flexibility of gender stereotyping in White upper-middle class adolescents in Grades 4

through 11. Their research results indicated that the propensity for gender stereotypes

concerning the psychological attributes of men and women increased during and

immediately after the transition from elementary to junior high school. After the

transition was complete, and over the remaining years of junior and senior high school,

stereotype flexibility increased.

African American girls may have an advantage in fending off negative gender

stereotypes: Black families generally do not distinguish between characteristics that are

considered appropriate for males and those that are considered appropriate for females,

and both boys and girls are socialized to be independent and to achieve (Scott-Jones &

Clark, 1996). Other research indicates that African American females of all ability levels

do not seem to endorse stereotypes of male academic subject domains. Ford (1992) and

Yong (1992), for example, found that gifted female African American middle school








students were just as positive in their attitudes toward math and science as their male

counterparts.

These research studies suggest that observers should guard against making

generalizations about academic achievement based on race or ethnicity. Studies on racial

and ethnic differences in academic achievement indicate that there are notable differences

among students of color that defy categorization (Steinberg, Dornbusch, & Brown, 1992).

There seems to be no single protective factor responsible for positive academic

achievement that can be named for young women, but resilience is likely a combination

of factors including resistance to gender stereotypes, resistance to racial and class

stereotypes, high self-esteem, and good use of coping skills.

In summary, adolescence is a time of tremendous change that provides a ready

context for factors of risk to develop for female adolescents. The process of gender

intensification serves to exasperate gender-role stereotypes that inhibit girls' academic

achievement. Protective factors may include the willingness to speak up for oneself.

Resisters, girls who resist gender-role and other types of stereotypes, learn to read social

messages, scan them for acceptance or rejection, and react accordingly (Stevenson, Reed,

Bodison, & Bishop, 1997).

Academic Achievement

In this section, research studies and theory will be discussed that document the

decline in female academic achievement. Focus will be given to studies that investigate

the question of when academic decline occurs and the differences between standardized

test scores and classroom grade decline. This will be followed by a discussion about

academic achievement and gender-role socialization. The section will conclude with a








description of the middle school environment and differential treatment in schools of girls

and boys.

Decline in Academic Achievement

Research on the topic of adolescent females' declining academic achievement in

middle and high school is sparse with the exception of occasional reports about girls and

mathematics. Public interest was generated by the AAUW (1991, 1992) reports, the

Sadkers' (1994) book, Failing at fairness: How America' schools cheat girls, and Pipher's

(1994) Reviving Ophelia, but did not extend beyond a few years at best. More scholarly

research on the topic is needed (Reis & Callahan, 1996).

What is known is that girls' academic achievement makes a decline sometime

between school entry at kindergarten and the completion of high school (AAUW, 1992).

Although the decline is not uniform across race, ethnicity, social class, or geographic

location, the decline is significant and widespread, and has captured the attention of

researchers who study this area (Orenstein, 1994). In their compelling study of girls and

women in American schools, the Sadkers (1994) wrote, "Females are the only group in

America to begin school testing ahead and leave having fallen behind. The longer girls

stay in school, the further behind they fall, especially in the areas of mathematics and

science" (p. 138).

For the many young women who experience declining academic achievement, the

consequences are often serious and life-long (Benbow & Stanley, 1983; Crockett &

Crouter, 1995). Young women receive a disproportionate amount of low standardized

test scores that prevent them from attending the country's best colleges and graduate

schools (Sadker & Sadker, 1994).








In this study a distinction will be made between academic achievement as

measured through standardized test scores and as measured through classroom grades.

Although both means of assessment are important and utilized frequently, standardized

test scores are used in additional ways, such as applying for college scholarships, that

grades are not (Sadker & Sadker, 1994).

Many females get good grades in most academic subjects throughout elementary,

middle, and high school (Wentzel, 1988). Good grades may result from the tendency for

classroom grades to reflect social competence, social skills, and positive attitudes towards

compliance. Some observers belief that girls do well in school because they are socialized

to follow instructions from teachers and accept rules with little protest (Benbow & Minor,

1986; Gilligan et al., 1990; Lindley & Keithley, 1991; Mickelson, 1989).

There are a few research studies that help to locate when the decline in female

academic achievement begins. In a longitudinal study of 30 males and 30 females from

sixth through twelvth grade, Wentzel (1988) examined math and English achievement.

Measuring achievement by both classroom grades and standardized test scores, her

results indicated that female math and English classroom grades remained stable over

time but achievement test scores in both subjects declined steadily for females compared

to males over the same period of time (Wentzel, 1988). Other longitudinal studies of

student performance indicate similar results (Lewis & Hoover, 1983).

In another effort to pinpoint when the decline in girls' achievement begins, results

from a comprehensive national assessment measure can be examined. The National

Assessment of Educational Progress (NAEP) is used in many states as a measure of

achievement in several different subjects in fourth, eighth, and eleventh grades. The








NAEP points to across-the-board female loss in achievement, with records indicating that

in elementary school girls outperform boys in every subject excluding science. However,

girls' NAEP test scores begin to descend in middle school and by high school boys'

scores are greater than girls' scores in every subject including math and science (Hafner,

1990).

The results of these research studies and others indicate fairly conclusively that

there is indeed a decline in girls' grades and test scores and that it generally begins during

the middle school years. This conclusion has been found on the international level as

well. The Third International Mathematics and Science Study, one of the largest studies

ever undertaken regarding educational achievement, collected test scores from 45

countries. Although data were not available from elementary school grades, the study

results showed seventh and eighth grade boys outperforming girls in almost every one of

the participating countries (Beaton, 1996).

Many of the studies report differing academic achievement results according to

the academic subject area examined. Part of the reason for the disparate results may be

the many different research methodologies used in the various studies. Scholars have

pointed out the need for clarification and consistency in research designs that investigate

academic achievement for girls and women, citing the oftentimes conflicting results of

published research (Reis, 1991).

Scores on standardized tests diverge the most between adolescent females and

males in the subjects of math and science. Some studies show that males tend to

outperform females on standardized tests of mathematics problem solving, but that girls

obtain higher grades in mathematics than boys do (Kimball, 1989). Some research








indicates that gender differences are specific to the particular aspect of math skills being

examined (e.g., math computation or problem solving) and that boys do better in only

some aspects of mathematics than girls (Lummis & Stevenson, 1990).

Hyde, Fenneam & Lamon (1990) found in their meta-analysis that gender

differences in mathematics performances were greater for White Americans than for

Black, Hispanic, or Asian Americans. Gender differences favored males in mathematical

problem solving in high school and through college. About 43 % of high school females

but 57 % of high school males scored above the average score for the whole high school

sample. It is important to note that mathematical problem-solving skills are critical for

success in such fields as engineering and physics (Lips, 1993).

Gender differences in science achievement continue to favor males over females,

and have shown little change for the past two decades. Boys continue to attain higher

scores than girls on national assessments of science achievement at the elementary,

middle school and secondary levels (Mullis et al., 1994).

Differences between male and female achievement levels among girls of color

and girls of differing socioeconomic backgrounds vary considerable. Ethnic minority

girls have significantly lower achievement scores and take fewer advanced courses in

mathematics and science than their White peers. These achievement gaps appear early in

development and tend to increase with age (Meece, 1996).

Differences in academic achievement impact students of all abilities. Among

gifted students the differences between test scores for males and females is even greater

than among nongifted students (Rosser, 1989). Gender differences between high ability








males and females in math and science is as large today as it was twenty years ago

(Hedges & Nowell, 1995).

Geographic location may also be a factor in academic achievement. In 1996, for

example, the state of Florida participated in the previously mentioned National

Assessment of Educational Progress (NAEP). In Florida, 2,353 students from 105 public

schools were assessed in a number of subjects including earth, physical, and life sciences.

The average proficiency in Florida compared to other states in the United States was

average and the average science score of males did not differ significantly from that of

females in Florida. However, White students in Florida had an average science score that

was higher than those of both Black and Hispanic students (O'Sullivan, Jerry, Ballator, &

Herr, 1997).

By the time students complete high school the gender gap has opened wide. Many

studies have focused on college-bound students, since these students participate in a

number of standardized tests including the Preliminary Scholastic Assessment Test, or

PSAT, that is taken in the junior year of high school. The PSAT is used by students to

apply for college scholarships. Boys' scores on the PSAT are so much higher than girls

that two out of three National Merit Scholarships semifinalists are male (Sadker &

Sadker, 1994). Eight thousand girls score among the highest category of the PSAT while

18,000 boys reach this level, with boys outscoring girls on both the verbal and the math

sections of the PSAT (Mullis, 1991).

On the SAT, boys on the average score approximately 60 points higher than do

girls. Males typically outscore females by approximately 50 points on the math section

and by 10 points on the verbal section (Connor & Vargyas, 1992; Hyde, Fennema, &








Lamon, 1990; Sadker & Sadker, 1994). Females of color do not score as highly on the

SAT as do White students, with the exception of Asian-American students. Within each

racial group, minority girls consistently score between 19 to 62 points lower than

minority boys do (Connor & Vargyas, 1992; Sadker, & Sadker, 1994).

Academic Achievement and Gender-Role Socialization

In 1992 Mattel unveiled a new Teen Talk Barbie. When the popular doll hit the

stores, one of the doll's comments was, 'Math class is tough' (Sadker & Sadker, 1994).

An important aspect to understanding the drop in academic achievement observed

in young women is observation of the link between academics and gender-role

socialization. Societal sex role stereotypes continue to exert a tremendous influence on

the academic direction of young women, leading adolescent girls to believe that

achievement is a masculine endeavor (Hyde, Fennema, Ryan, et al., 1990; Wentzel,

1988). Many young women are afraid to look too intelligent because they fear they will

face social rejection by boys who either do not value intellectual ability as a feminine

attribute or may feel threatened lest they feel less competent by comparison (Orenstein,

1994). Smart girls may feel alienated from their girlfriends as well who may view them

as show-offs or too academically competitive. Rather than risk social censure, some girls

choose not to express their knowledge or opinions, particularly within the public

classroom setting (Harter et al., 1997).

Math and science courses are widely perceived by both boys and girls as the more

masculine of academic realms. For example, in a recent study in England and Wales 342

high school females and males were asked their perception of school subjects in

relationship to gender appropriateness. Study results indicated that science was seen as a








masculine subject by both girls and boys, and arts and languages as feminine subjects.

Both sexes admitted, however, that both boys and girls are equally good at most of their

academic subjects (Whitehead, 1996). Young women who perceive math as a male

endeavor are more likely to avoid math courses in high school than girls who do not

(Eccles, Barber, & Jozefowica, 1999; Hyde, Fennema, & Lamon, 1990). By opting out of

math and science courses, girls may find their access to careers in science and technology

severely limited (Wigfield & Eccles, 1994a).

Girls who view mathematics as a masculine subject tend to do more poorly in

their academic achievement than girls who do not hold that view (Kloosterman, 1990;

Meyer & Koeher, 1990). A key finding in the AAUW (1992) report is that girls' interest

in math and science drops dramatically as they advance through school. Even girls who

like the subjects are, by age fifteen, only half as likely as boys to feel competent in them.

These findings are key because educators find that a loss of confidence in math usually

precedes a drop in achievement, rather than vice versa (AAUW, 1992; Lapointe, Mead,

& Phillips, 1989).

In the area of natural science, middle school girls have similarly been found to

self-select themselves out of advanced courses. A two-year longitudinal study of 155

adolescents in seventh and eighth grades was conducted to assess the developmental

relationship of perceptions of self-concept and gender role identification with

adolescents' attitudes and achievement in science. Results indicated that students' self-

concept and gender role perceptions were related to both achievement and attitudes

toward science, but more related to attitude than achievement. Girls who thought that

science was a more masculine endeavor continued to achieve in science, but their attitude








toward science became more stereotypical as they progressed from seventh to eighth

grade (Handley & Morse, 1984).

Adolescent females create covert strategies for not appearing too bright. Sadker

and Sadker (1994) spoke with girls who confided that some of their strategies for

minimizing their academic participation included taking inconspicuous seats in the back

of the room, checking where the teacher never looked and then sitting there, and only

raising their hands halfway.

Some girls are more reticent to adopt stereotypical gender-role beliefs than others,

andtheir resistance appears to have a positive impact on their academic performance. For

example, a recent study of 67 families with young adolescents was conducted with the

purpose of exploring implications of parents' traditional versus egalitarian marital roles

and the relationship with their daughters' and sons' math and science achievement.

Findings revealed that young women from families that practiced a more egalitarian

division of household tasks maintained a higher level of achievement across the transition

to seventh grade, but females from traditional families declined in their math and science

performance (Updegraff, McHale, & Crouter, 1996). These results suggest that girls who

have less stereotypical role models in their own mothers will be more likely to resist

gender-role stereotypes and to demonstrate more confidence in achievement situations

(Jacobs & Eccles, 1992).

Middle School

The classroom environment in middle school plays a significant role in the

academic achievement of female adolescents, at times having a negative impact. The

classroom environment becomes less personal, more competitive, and more ability-








centered as young people progress through school, and these changes are especially

evident during the move from elementary school to middle school (Eccles & Blumenfeld,

1985). Occurring at a time of heightened self-consciousness for adolescents, the

transition is particularly problematic for girls who find their social networks disrupted.

There is a decrease in opportunity for close student-teacher relationships (Wigfield &

Eccles, 1994b). Because of the increased emphasis in middle school on individualism,

changes in the nature of the classroom environment are less compatible with norms for

femininity and more compatible with norms for masculinity (Raeff, 1997; Roberts,

Sarigiani, Petersen, & Newman, 1990)

Reports from adolescent females of difficult middle school experiences have been

corroborated by the girls' teachers. Taylor, Gilligan & Sullivan (1995) interviewed a

number of middle school teachers for their Understanding Adolescence Study. The

teachers cited a lack of connection with both colleagues and students in their urban

middle and high schools and spoke of overextended workloads, unexpected lay-offs, and

random placement in different schools each year that discouraged personal relationships

both with students and colleagues.

Differential Treatment in School

In the past decade several prominent studies have been published suggesting that

girls and boys are not treated equally in the classroom and, as a consequence, have an

unequal chance to succeed in higher education.

One of the most well publicized studies on gender, academic achievement, and

self-esteem was commissioned by the American Association of University Women and

published in 1992. Titled, Shortchanging Girls, Shortchanging American: How Schools








Fail Girls, the results from this study indicate a state of unequal practices and

consequences resulting in the erosion of girls' academic self-esteem and academic

achievement. The AAUW study involved over 3,000 boys and girls between the ages of 9

and 15 across the United States. The report suggested that classroom practices of

teachers, textbooks, tests, and policy often contribute to the inequitable treatment of the

sexes in schools. In their research the AAUW found that boys received more attention in

school through a number of different ways. Boys are called on more often in the

classroom than girls and are asked more abstract, open-ended and complex questions, and

they are twelve times as likely to speak up in class as girls. Teachers choose activities in

their classrooms that appeal more to boys than to girls, and they praise boys more for

their academic and intellectual work, while they praise girls more for their clothing and

good behavior. Textbooks used in public schools featured seven times as many

illustrations of boys as of girls, and boy-centered stories were used three times as often as

girl-centered stories. Adjectives in the stories used to describe boy characters included

clever, brave, creative and resourceful, while girls were most often depicted as kind,

dependent and docile. The typical school child reads six times as many biographies of

males as of females (AAUW, 1992).

Myra and David Sadker elaborated on the AAUW report (1991, 1992) in their

book, Failing at Fairness (1994). In their own research on the topic, they reported on a

content analysis of 15 math, language arts, and history textbooks currently used by school

districts. Numerous examples of gender stereotyping were found. For example, a 1989

sixth grade history textbook had four times as many males' pictures as females and only








11 female's names were mentioned in the 631-page textbook (Sadker & Sadker, 1994).

Other studies have documented similar findings (Siegle & Reis, 1995).

In investigations of teacher behavior, a number of researchers have reported that

teacher bias generally favors males over females, and middle and upper-class students

over low socioeconomic class students (LaFrance, 1991; Takei, 1993). Considered a

reflection of society, these types of teacher practices nonetheless have deleterious impact

on students in a myriad of ways including academic achievement (Alexander, Entwisle,

& Thompson, 1987; Sadker & Sadker, 1994). Gender stereotypes are conveyed to

students both overtly and covertly and create inequities in the classroom that restrict the

potential of students to learn and achieve (Lindley & Keithley, 1991). For example, in her

analysis of the role of teacher beliefs on mathematics performance, Fennema (1990)

reported that teachers perceive that ability explains their most capable males' success

58% of the time, but explains their best females' success just 33% of the time.

In conclusion, the research and theory reports presented here illustrate that

many girls today still believe that being bright is in conflict with being popular. Gender

stereotypes that equate masculinity with high academic success and that discourage

females from pursuing academic achievement pose a widespread risk for girls in middle

and high school. Many girls opt out of challenging academic pursuits, especially

advanced mathematics and science. Resisting these gender stereotypes can be a powerful

protective factor for adolescent females.

Self-Esteem

In this section research and theory that addresses the role of low self-esteem as a

risk factor for adolescent females and high self-esteem as a protective factor will be








investigated. Studies will be highlighted that shed light on the following questions: Does

self-esteem drop for girls in middle school? How does ethnicity impact self-esteem? Is

there a relationship between self-esteem and femininity? Is there a relationship between

self-esteem and coping skills?

Healthy self-esteem can be seen as a protective factor for young women. Female

adolescents with healthy self-esteem tend to have an appropriate sense of their potential,

their competence, and their innate value as individuals (Callahan, Cunningham, &

Plucker, 1994). Adolescent females with low self-esteem tend to deride their own

abilities and discount their successes. Poor self-evaluation leaves adolescent females

looking outward for approval and leaves young women particularly vulnerable to putting

their self-esteem in the hands of boyfriends (Callahan et al., 1994; Cohen, 1995;

Orenstein, 1994).

The decline in girls' self-esteem as they enter adolescence has been termed an

"epidemic" (Pipher, 1994). The American Association of University Women's survey

(1992) reports that in response to the statement, "I'm happy the way I am," 60 % of the

girls polled responded affirmatively in elementary school, but by the time they reached

middle school, only 37% responded affirmatively, and only 29 % did so in high school.

The self-esteem of Hispanic girls fell the most, with a total drop of 38 percentage points

between elementary and high school. In interviews students reported that girls are quieter

in the classroom than the boys. Almost half of the boys in middle school said they speak

up in class, but only 39 % of the girls agreed that they spoke up in class. The exception to

the plummeting report of self-esteem was demonstrated by African American females.








Their positive sense of self-worth, already higher than other girls in elementary school,

held steady in middle school, and actually increased to a high of 74% by high school.

In attempting to pinpoint the decline in girls' self-esteem, researchers have

employed longitudinal studies to try to determine the timing and duration of the plummet.

Results consistently find that as girls move toward adolescence their self-esteem drops in

relation to their own level of self-esteem at a younger age, as well as in contrast to the

self-esteem of their male peers. An associated loss frequently seen concurrently in young

women includes the loss of optimism about their lives (Brown & Gilligan, 1992; Surrey,

1991).

Some studies have targeted the middle school period directly for an examination

of self-esteem. Wigfield & Eccles (1994b) surveyed 1,850 low- to middle-income White

students to learn more about the way that the transition from elementary to junior high

school influences children's self-esteem. Students completed questionnaires at four

different times: twice in the sixth grade and twice in the seventh grade. The authors

reported that all of the children's self-esteem decreased significantly across the transition

to seventh grade, but that boys reported higher self-esteem than did girls at all four

assessment points. In contrast, the authors found no differences in boys' and girls' self-

esteem levels in elementary school. These results lend support to other research findings

that indicate that girls' self-esteem does decline during the middle school years.

There are various explanations offered for the sex-differentiated decline in self-

esteem. Rosenberg (1986) suggested that girls are more affected by the physical changes

that occur at puberty and that these changes impact females' self-esteem. Other

explanations focus on the environmental changes that accompany the transition to middle








school, and point out that the emphasis on competition, social comparison, and ability

self-assessment is particularly difficult for girls when coupled with their heightened self-

focus (Wigfield & Eccles, 1994b).

One of the earliest studies to investigate whether any connection existed between

gender role stereotypes and self-esteem in early adolescence was undertaken by Lamke in

1982. In her research she assessed 119 junior high students using the Rosenberg self-

esteem measure, the Ber Sex Role Inventory, and the Personal Attributes Questionnaire.

She found that girls or boys with high levels of instrumental traits (masculine) had higher

levels of self-esteem than feminine and undifferentiated individuals. Expressive traits

were not associated significantly with self-esteem, thus indicating a link between

masculine traits and high self-esteem. Another more recent study that found similar

results was reported by Rose and Montemayor (1994) who studied 194 White adolescents

in the sixth through twelfth grades. In this study girls who scored highest on masculinity

had the highest levels of self-esteem as well as the highest levels of perceived scholastic

competency. No grade differences were found in these relationships.

Self-esteem has been shown to affect coping skills in adolescents. Successful

coping that allows the individual to face and resolve conflicts leads to increased self-

confidence and self-esteem (Bednar, Wells, & Peterson, 1989). In research with 417

ethnically diverse, urban high schools students, Kimbauer (1992) found that the variables

of self-esteem and ethnic identity worked together to significantly affect the coping

strategies used by the adolescents. Kimbauer found that when ethnic minority adolescents

were confronted with racial stereotypes and/or prejudice, their choice of coping strategies

was related to both their self-esteem level and their sense of ethnic identity. Neither








variable when examined independently was significantly related to the type of coping

strategy used. Those adolescents with both high self-esteem and high ethnic identity

tended to use assertive coping strategies, such as talking with pride about their ethnic

group or engaging in direct discussion of the issues at hand.

A large number of research studies supports the existence of a relationship

between academic achievement and self-esteem (Muijs, 1997). Research that targets the

middle school years consistently documents significant declines in both self-esteem and

academic achievement for girls (American Association of University Women, 1992;

Rothenberg, 1995). Yet despite the many studies showing positive correlations between

these two factors, there is little agreement as to either the reasons behind the decline or

the causality of the two variables. A review of the research studies in this area reveals

studies that indicate that self-esteem enhances academic achievement, and just as readily,

other studies that indicate that academic achievement enhances self-esteem (Muijs,

1997).

In research examining self-esteem and academic achievement for fifth- and sixth-

grade students, data were collected from a self-esteem assessment and from results of the

Iowa Tests of Basic Skills. Researchers found significant correlations between self-

esteem, grade point averages, and Iowa Tests of Basic Skills test scores for both boys and

girls. Self-esteem scores were more predictive of grades than the composite score on the

ITBS (Wiggins, Schatz, & West, 1994).

Path analyses methodology has been used in some of the research studies

examining the relationship between academic achievement and self-esteem. Most path

analysis studies in this field have set out to unravel the causal chain (i.e., which comes








first, self-concept or achievement, or is the relationship reciprocal?). The results of this

effort, however, have been contradictory (Muijs, 1997).

One of the most striking aspects of the AAUW report (1992) was its finding of

the high self-esteem levels of African American female adolescents. Black girls in the

study began with and were better able to retain higher levels of self-esteem throughout

adolescence than their White and Latina counterparts. Differences in girls' self-esteem by

race and ethnicity have since been documented in other research studies as well. A 1996

study of young people in two urban high schools examined the relationship between

school context, ethnicity, grade level and self-esteem. While noting that girls had

significantly lower self-esteem than did boys, the authors reported that African American

female adolescents reported higher self-esteem levels than either White or Hispanic

students and that adolescents who rated themselves as strongly ethnically identified had

higher scores for self-esteem than those who self-identified as mainstream or bicultural

(Rotheram-Borus, Dopkins, Sabate, & Lightfoot, 1996).

The AAUW's report of self-esteem differences by ethnicity was not the first to

find high self-esteem in African American youth. In 1971, Morris Rosenberg proposed

that African American children had higher or equivalent self-esteem than did White

children. Since them, numerous other studies have replicated Rosenberg's findings

(Martinez & Dukes, 1991; Rosenberg & Simmons, 1971). Of particular interest is the

finding that higher self-esteem levels among Black children relative to Whites is present

at all economic levels (Hunt & Hunt, 1977).

Research reports are mixed in regard to the relationship between self-esteem and








academic achievement among African American youth. One of the more recent studies,

utilizing a sample from Gainesville, Florida, examined the relationship of self-esteem to

academic achievement in 59 African American children from low-income families.

Ranging from 1st to 12h grades, findings indicated that self-esteem was not a significant

indicator of academic achievement, grade point average, or reading or math scores on the

California Achievement Test for either boys or girls, and that these findings were

consistent across all grade levels (Gaskin-Butler & Tucker, 1995).

Despite the widespread utilization of self-esteem research, there has been some

criticism of the concept of self-esteem as typically used in the behavioral science

research. Judith Jordan (1991), for example, states that the traditional notion of self-

esteem is derived from a separation model rather than a relational model, and thus tends

to favor males. Viewed from this perspective, healthy self-esteem is construed to be

comparison based, implying that the individual feels good about herself in comparison

with others, and more specifically, if she feels better than she perceives others do. Jordan

contends that a more relevant construct of self-esteem for girls and women is one that is

based on self-in-relation theory (Miller, 1991; Surrey, 1991) and that describes a mutual

self-esteem based on emotional sharing and shared sense of understanding and regard

(Jordan, 1995; Surrey, 1991).

The paradigm of self-esteem as a measurement of how good one feels about

oneself from what Jordan (1991) describes as a comparative or independent perspective

remains, however, the model of choice for the vast majority of researchers examining the

topic of resiliency. Many of the studies in the field of resilience utilize the global self-

worth construct operationalized by Rosenberg in his self-esteem measure (1965).








In summary, findings from previous research results provide evidence that the

self-esteem of most adolescent females does drop during the middle school years.

Research findings report a higher level of self-esteem for African American than White

female adolescents, and some research studies indicate that there is a relationship

between low self-esteem and high femininity, and between high self-esteem and low

femininity.

Resilience

This study examined the theory of resiliency for its ability to explain the

phenomenon of female adolescents' declining academic achievement, to gain a better

understanding of the problem, and to increase the chances that effective interventions can

be developed and implemented. In this section the findings from research studies and

theory that illustrate the major constructs of risk and protective in relationship to

academic achievement are presented. Examples from the stress and coping literature are

included in the discussion of protective factors because in the resilience literature coping

is considered one resource available as a protective factor. The mechanisms by which

these factors work will be briefly examined.

Risk Factors

Poverty has been demonstrated to be one of the risk factors that consistently has a

detrimental impact for young people in a number of areas including academic

achievement. Many studies have found that poverty has in fact a tremendous impact on

children's educational achievement and cognitive development (Boals, Tyree, & Barker,

1990; Montgomery et al). As of 1994, 22% of American children lived in families with

incomes below the poverty threshold (U.S. Bureau of the Census, 1996). Children from








low socioeconomic backgrounds tend to score significantly lower than nonpoor and

middle-class children on numerous indicators of academic achievement, including

achievement test scores (Conger, Conger, & Elder, 1997). Meta-analysis suggests that

family income is the highest single correlate of academic achievement, followed by

parental occupation and parental education (McLoyd, 1998). Persistent poverty is

repeatedly found to have more detrimental effects on IQ, school achievement, and

socioemotional functioning than does occasional or transitory poverty (Cornwell, 1993;

McLoyd, 1998).

Of the many studies that have documented the detrimental effect that poverty has

on children's academic achievement, a study by Walker, Greenwood, Hart and Carter

(1994) specifically investigated the impact of low family income on standardized

achievement test scores. In this longitudinal study the authors found that children from

low-income homes obtained lower scores on standardized reading and spelling

achievement tests throughout their elementary school years.

McLoyd (1997) summarized numerous studies of children and adolescents that

report an association between socioeconomic disadvantage and socioemotional problems

with samples of White and African American adolescents and children. McLoyd noted

that low socioeconomic class during early and middle childhood is associated with

diminished self-confidence and self-esteem as well as poor academic achievement.

An example of additional hardships facing low socioeconomic class students was

described in research by Branglinger (1991). In this study adolescents' reports of

problems and punishment in school were examined. Poor students, compared to their








more affluent schoolmates, reported a greater number of penalties, more severe,

stigmatizing punishment, and more stringent consequences for similar infractions.

A second factor that is often cited in resilience literature as a risk for children and

adolescents is exposure to major stressful life events. Research findings indicate that

some stressful events including parental divorce, death or serious illness of a family

member (Hashima & Amato, 1994) and community violence (Richters & Martinez,

1993) have been demonstrated to have an association with lower academic achievement.

Stress is often a cumulative experience. Low family income and stressful life events are

two stressors that frequently occur simultaneously. Children raised in families with low

incomes are more likely to experience other risk factors in addition to poverty (DuBois,

1994; Pungello et al., 1996).

A third risk factor commonly cited in resilience literature related to the

maintenance of academic achievement is differential schooling. Schools in the United

States have different instructional practices and educational resources for poor and

affluent students. Teachers may exacerbate the divisions between the haves and the have-

nots in relation to the achievement gap by endorsing middle-class attitudes and values in

the classroom that clash or seem threatening to a student of low economic status

(Alexander et al., 1987; Takei, 1993). Poor children and adolescents frequently have a

higher prevalence of emotional and behavioral problems than their middle-class

counterparts (Adams, Hillman, & Gaydos, 1994).

Protective Factors

A number of research studies in the resilience literature have focused on

identifying those protective factors that some African American youth utilize to








overcome the disadvantages of growing up in a risk-filled environment that may include

racism, poverty, and poor educational opportunities. An example of a protective factor

utilized by African Americans is racial socialization, described as a proactive orientation

toward racial barriers that African American families teach their children and that focuses

on ethnic pride and egalitarianism (Ward, 1996). Racial socialization has been

demonstrated to have a positive effect on school performance among African American

high school students. Bowman and Howard (1985) conducted a study in which they

interviewed 377 African American adolescents between the ages of 14 and 24 years.

They found that those youth whose parents transmitted some message to them about the

reality of racism and the importance of ethnic pride and self-development had a greater

sense of personal efficacy and had higher grades in school. Racial socialization typically

includes parents teaching their daughters a healthy resistance to gender as well as racial

oppression (Ward, 1996).

Specific protective factors have been identified in both White and Black youth

who do not succumb to environmental risks. In a study of 442 high school seniors,

resilient youth were found to possess the following protective factors: an internal locus of

control, an active orientation to life, a tendency to seek out others, and the inclination to

follow their own preferences instead of giving in to peer pressure (Blocker & Copeland,

1994).

A particularly intriguing protective factor present for some adolescent females has

been identified by Carol Gilligan (1991) as political resistance. Gilligan explains that this

trait is manifest by young women who take action against social or cultural conventions

that serve to oppress themselves or others. Political resistance may include specific








actions that serve to reject stereotypes based on race, ethnicity, class or sex (Taylor et al.,

1995). Pastor, McCormick and Fine (1996) elaborate on this concept, contending that

some urban female adolescents of color develop social consciousness and actively reject

White middle-class values. Political resistance can be thought of as part of a successful

coping strategy.

Utilization of healthy coping skills is an integral facet of resiliency's protective

factors. Healthy coping strategies can mediate or alleviate the effects of stress (Lazarus &

Folkman, 1984). Investigations of the coping processes of adolescents is particularly

important because acquiring coping skills, like all endeavors for youth, is a

developmental process (Hauser et al., 1991). It may be that young people confront many

different types of life stressors that they have not encountered previously, and/or they

may not yet have a wide variety of coping strategies to rely upon (Patterson &

McCubbin, 1987).

McCubbin and Patterson (1982, 1983) have developed a theoretical model of

coping that takes a developmental approach to the acquisition and utilization of coping in

children and adolescence. Their theory integrates previous individual coping theory with

family stress theory. Coping is viewed as one of many resources that can add to or

decrease the amount of stress that adolescents experience; young people generally are

dealing with several demands simultaneously. McCubbin and Patterson (1983, 1985)

believe that adolescents must learn how to manage individual, family, and community

demands, and that coping is successful when the adolescent is simultaneously able to fit

into his or her family and community. Bandura (1981) also noted the complexity

involved in adolescent coping, noting that successful adolescent coping must include








learning how to coordinate cognitive, social, and behavioral skills to deal with stressors

that are often ambiguous and unpredictable (Bandura, 1981).

Some speculation has gone into examining factors that affect the development of

adolescent coping skills. Some researchers have examined whether the timing of puberty

affects coping, with results suggesting that timing has little impact on coping for either

sex (Bush & Simmons, 1987). Nolen-Hoeksema (1995) suggested that gender

differences in coping style likely result from gender role stereotypes that emphasize that

males are active and ignore their feelings, whereas females are passive and emotional.

Gender-typed socialization increases the chances that adolescent females will adopt

helpless, passive coping behaviors.

Studies have documented the differences in stress and coping between male and

female adolescents including the type and frequency of use of various coping strategies

(Copeland & Hess, 1995). Bird and Harris (1990) found that young adolescent females

reported feeling more strain in their family role than their male counterparts, and that

they tended to manage life problems by relying on social support (e.g. crying, talking to a

friend) and increasing involvement in interpersonal relationships with friends, siblings,

parents, and other adults. Girls may also participate in less healthy ways of coping with

anxiety or insecurity by using strategies such as avoiding negative reactions or being

compliant (Hill & Lynch, 1983).

The importance of ethnicity as a potential variable in the coping process cannot be

overlooked because ethnicity affects adolescents' perception of, and reaction to, stressful

life events (Phinney, Lochner, & Murphy, 1990). Spencer, Cole, DuPree, Glymph, &

Pierre (1993) found from their research with 562 African American urban middle school








adolescents that African American adolescents who have high self-esteem and good

coping skills tend to demonstrate higher academic achievement in school. Spencer,

Dobbs, & Swanson (1988) report that minority adolescents who identify with their own

ethnic group have greater resilience during periods of unusual stress, including better

mental health and academic performance.

Girls of color may learn important coping mechanisms that enable them to filter

out degrading racial and gender messages (Rotheram-Borus, Dopkins, Sabate, &

Lightfoot, 1996). Robinson and Ward (1991) suggest that African American females who

fend off negative stereotypes about themselves often have higher self-esteem and other

healthy attitudes and behaviors than females who succumb to stereotypes.

In conceptualizing resilience in urban youth growing up in high-risk

environments, Spencer et al. (1993) propose that self-efficacy regarding academic

achievement can be conceptualized as a method of coping. Academic self-esteem is

viewed as a coping method because educational attainment remains the most important

means for urban youth to escape chronic poverty. In their longitudinal study of 562

African American adolescents, the researchers examined coping methods and

competence outcomes as measured by scores on the Iowa Test of Basic Skills and on a

measure of academic self-esteem. With resilience conceptualized as academic coping,

regression analyses indicated that significant predictors of academic performance for

females were mother's education, parental life dissatisfaction, youth's perception of

family conflict, and academic self-esteem. The findings indicated that as adaptive coping

responses, both academic self-esteem and academic achievement are responsive to

particular protective factors. The researchers also found that elementary school girls








obtained higher academic achievement scores than middle or high school females, and

that a decline in academic achievement was observed in girls' middle school performance

(Spencer et al, 1993).

Taken together, previous research and theory building literature indicates that

resiliency theory lends itself to a description of stereotypical gender-role beliefs as both a

potential risk or protective factor for adolescent females. Based on the resiliency

literature, it was expected that low socioeconomic class would be associated with a

decline in adolescent females' self-esteem, academic achievement, and coping skills.

Adolescent females who have healthy coping skills, high self-esteem, and who are able to

resist gender role socialization messages were expected to more likely maintain academic

achievement than young women with poor coping skills, low self-esteem and adherence

to stereotypical gender-role socialization messages.













CHAPTER III
METHODOLOGY

Overview

The purpose of this study was to examine the relationships among adolescent

females' academic achievement in middle school and traits of resilience in young

women, including high self-esteem, healthy coping skills, and low adherence to

stereotypical gender role beliefs. A within-subjects correlational design for survey

research was utilized for this research. Data were gathered on self-esteem as measured by

the Rosenberg Self-Esteem scale, on the utilization of coping skills as measured by the

Adolescent Coping Orientation for Problem Experiences (A-COPE), on adherence to

stereotypical gender role beliefs as measured by high femininity or high masculinity on

the Personal Attribute Questionnaire (PAQ), and on academic achievement as measured

by comparing the participants' fifth and eighth grade test scores on standardized tests and

their fifth and eighth grade classroom grades. A demographic component provided by the

various school systems was used to provide information on participants' ethnicity, age,

and socioeconomic status.

This chapter describes the methodology employed in this study. Included are

descriptions of the population, sample, sampling procedures, research design,

instrumentation, and data analysis. The chapter concludes with a discussion of the

methodological limitations of the study.








Population

The population from which the sample for this study was drawn is ninth grade

students attending public high schools in Alachua, Bradford, Clay, and Putnam counties

in the state of Florida. Since standardized tests are generally given to students in April of

eighth grade, and scores are not available until the end of the academic school year, data

for this study were collected during the first part of the 1998- 1999 school year from

incoming ninth grade students.

Student participants included residents of small or medium sized cities and rural

areas. The geographical area of Alachua County, for example, is fairly evenly divided

between Gainesville, a university city of 100,300 residents in the north central portion of

Florida, and the more rural areas of the county that surround the city. The average age of

participants was 14.5 years. The demographic composition of the schools involved is

fairly typical of many Southern states. At Hawthorne High School, for example, the

1998 -1999 school year demographics for the total study body of 650 students is 340

males, 310 females, 409 White, 231 African American, 5 Hispanic, 3 Asian, 1 Native

American and 1 Multiracial. Students from low socioeconomic background are well-

represented in this geographical area. Forty-four percent of students in Alachua County's

middle schools, for example, are eligible to receive free or reduced lunches; the state-

subsidized lunch program is based on household income, with a family of two reporting a

yearly income below $14,000 (Alachua County School Board, 1998-1999).

Sample and Sampling Procedures

Permission was received from the University of Florida Institutional Review

Board for this research. An application for permission to do research in the schools was








submitted to the School Boards' Department of Research and Evaluation and permission

received for each of the four participating counties as well as from P.K. Yonge

Developmental Research School of the University of Florida.

The sampling plan used in this research was a convenience sample. After each site

principal agreed to participate, the researcher contacted teachers at the site schools. The

researcher met personally with each teacher, explained the research procedures, and gave

Informed Parental Consent forms to each teacher to distribute to their students (see

Appendix B). Parent/guardian signatures were obtained and the forms returned prior to

the day of data collection. During the meeting with the teachers, the researcher arranged

the time, date, and location for data collection. The researcher presented an incentive to

the teachers in exchange for granting access to their students An offer was made in

which the researcher conducted a class or workshop on stress management for high

school students that was scheduled at the conclusion of the survey administration.

The researcher asked the teachers for permission to enter their regularly scheduled

class to administer the questionnaire to all students in their class (es) who volunteered to

participate. When data collection commenced, the assessment explanation, survey

administration, and debriefing took approximately 20 minutes to complete.

This study required a minimum sample size of at least 200 adolescent females. In

the use of path analysis, the minimal sample size needed to provide adequate estimates of

the regression coefficients is 200 cases (Shavelson, 1988). However, because logistics of

working in the public schools necessitated that the survey be administered to both girls

and boys, it was necessary to recruit more students to complete the assessment measure,

although the data collected from male students was not be used in this research study. In








order to analyze the data by race and socioeconomic class, a minimum of 30 respondents

for each category was needed. Given the demographics of north central Florida's public

schools, enough data were generated to successfully analyze the findings for White and

African American females, and for lower income and higher income students. In order to

ensure an adequate sample size and to guard against incomplete data such as transfer

students, the researcher solicited more than the minimum number of participants actually

needed for this study.

The data measuring academic achievement were obtained from student records

made available by the either the individual high school or the county school board's

Administrative Offices. Standardized test scores from fourth or fifth grade and eighth

grade, as well as the classroom grades for math and reading from fifth and eighth grades

were obtained for each participant in the research. Student records were used to obtain

demographic information that included gender, age, ethnicity, and enrollment in the free

or reduced lunch program.

Design

The design for this study was a within-subjects, correlational design for survey

research. The independent (predictor) variables were stereotypical gender role beliefs,

healthy coping skills, and self-esteem. The dependent (criterion) variable was academic

achievement, and more specifically, the maintenance of academic achievement from

elementary to middle school. The three independent variables were operationalized by

(a) the Personal Attribute Questionnaire (PAQ) that assessed stereotypical gender role

beliefs, (b) the Adolescent Coping Orientation for Problem-Experience (A-COPE), that

assessed healthy coping skills, and (c) the Rosenberg Self-esteem scale that assessed self-








esteem. The dependent variable, academic achievement, was operationalized via two

measures that were gathered from student records: the difference between fifth and eighth

grade standardized test scores, and the difference between fifth and eighth grade

classroom grades in math and reading.

Instrumentation

Predictor Variables

The Personal Attributes Questionnaire (PAQ; Spence, Helmreich, & Stapp, 1974)

was selected as a measure of sex role identity for several reasons that fulfill the objectives

of this research. First, the instrument measures masculinity and femininity only in terms

of positively valued characteristics. Second, the PAQ has two scales, "Self" and "Other,"

that differentiates the beliefs that one holds for oneself and for others. Because peer and

social influence is such a critical factor for adolescents, it was thought that the inclusion

of this scale could provide an additional source of pertinent data in this study. The third

and perhaps most important reason for using the PAQ in this study was in its

appropriateness for use with young adolescents. The PAQ has been demonstrated to have

a reading and comprehension level suitable for young adolescents (Murphy & Steven,

1987). Finally, the solid reliability and validity data on the PAQ includes research with

adolescent populations of diverse backgrounds (Spence, 1978).

Spence et al. (1974) developed the PAQ because of concerns about traditional

measurement of the social constructs of masculinity and femininity. The authors based

the development of this instrument on the premise that masculinity and femininity are

two different dimensions of a personality construct rather than continuous traits. By

viewing masculinity and femininity in this way, the researcher is able to view an








individual as having both masculine and feminine psychological characteristics. For

example, rather than assuming that aggressiveness is a masculine trait and nurturance is a

feminine trait, in the dualistic approach a man or a woman can be both aggressive and

nurturant at the same time, or can have neither of these characteristics.

Much of the research of the past several decades in the study of gender

phenomena has been accomplished by the use of one of the two most popular assessment

measures: the PAQ or the Bem Sex Role Inventory (BSRI; Bem, 1974). Although the two

personality inventories measure similar constructs, there are noteworthy differences

between the two. In highlighting the reasons that the PAQ was selected for use in this

research rather than the BSRI, those concerns most relevant to the current study are cited

here and include considerations that (a) the items on the PAQ are framed in terms of

positive characteristics, while the BSRI contains items than may be construed as either

positive or negative, (b) the PAQ takes a multifactorial approach to gender, whereas the

BSRI works from a unifactorial gender schema theory (Spence, 1993), (c) the PAQ has

an "Other" scale to measure what the respondent believes that others think about each

characteristic, but the BSRI does not; and (d) researchers have indicated that the language

of the BSRI is not suitable for young adolescents (Richmond, 1984).

Spence and Helmreich (1978) reported research results that indicated that their

dualistic conception of femininity and masculinity is valid for high school students of

diverse socioeconomic status and from differing geographic locations of the country.

Significant differences in self-esteem in both sexes for the four PAQ categories have

been found, with the highest self-esteem reported by high masculine/feminine students,

followed by the masculine, feminine, and undifferentiated, respectively. Helmreich








(1980) reported that samples from Mexico, Germany, Tunisia, and France appeared to

yield results similar to the norms found in American samples. Binion (1990) reported the

results of research in which the PAQ was administered to low- to middle-income African

American women with a median age of 23 years. Examining the degree to which self-

reported masculine and feminine characteristics dictate role preferences and gender

expected behavior, her research lends support for the ability of the PAQ to measure

personality traits that demonstrate endurance over time and situation across nonWhite

cultures.

The PAQ is a 55-item instrument that contains two major scales: one consisting of

socially desirable instrumental traits (e.g., independence, decisiveness) that are

stereotypically more characteristics of males than females (M scale), and the other

consisting of socially desirable expressive traits (e.g., tactfulness, awareness of others'

feelings) that are stereotypically more characteristic of females (F scale).

PAQ respondents are instructed to make two sets of ratings on traits that

stereotypically differentiate the sexes. The measure's instructions make no mention of

sex differences. First, the respondent indicates whether the trait listed is characteristic of

themselves on a 5-point Likert scale that ranges from not at all, to very. Second, the

same characteristics are listed again, and this time the respondent is asked whether that

trait is much more characteristic of a male or much more characteristic ofa female.

The PAQ produces four scores, including a "masculine" score, a "feminine"

score, and two scores that combine attributes considered both masculine and feminine

(Spence, et al., 1974). Scores range from 0 to 5 on both the Self scale and the Stereotype

scale, with the extreme male or female choice indicated by a score of 5.








Alpha coefficients were computed for the Stereotype and total Self scores as a

measure of internal consistency. The values were .01 and .90 for the men and women,

respectively, on the Stereotype scale and .73 and .91 for men and women on the Self

Scale. Alpha coefficients reported in Bionion's (1990) study found that the PAQ

femininity scale was r = .68 and the PAQ masculinity scale was r = .66.

A second analysis reported by the authors (Spence, 1978) that was related to

internal consistency involved part-whole correlations computed between each Self-item

and the subscale. The range of r's for the men and women respectively was .23 to .64,

and .24 to .70 for the male-valued items. For the female-valued items, the values ranged

from .22 to .56 and .27 to .55 for the men and women. Parallel values for the sex-specific

items were .23 to .61 and .19 to .64. All r's were significant. Although the means of the

men and women did not differ significantly for all Self items, within each sex the items

tended to correlate satisfactorily with the individual's masculinity or femininity as

reflected in their overall score on the subscale to which the items belonged.

Test-retest data from the authors' original research were with 31 subjects who

retook the PAQ after an interval of 13 weeks. The r's were .92 and .98 for men and

women respectively on the Stereotype scale, and .80 and .91 on the total Self Scale. The

values for the subscales varied from .65 to .91.

This research used the short form of the PAQ, which was developed by Spence,

Helmreich, and Stapp (1975), in conjunction with their longer form. The short form has

24 items and correlations with the full scale of 55 items are high: for the male-valued

subscale, the correlations on the total Self, full versus short is r = .92, for the stereotype

scale, full versus short is r = .94.








The Adolescent Coping Orientation for Problem Experiences (A-COPE; Patterson

& McCubbin, 1987) was selected for use in this research because of its unique focus on

the coping behaviors of adolescents. The measure was designed at a reading level

appropriate for young adolescents, and with items that realistically tap the range of

behaviors used by today's adolescents.

The A-COPE was developed by Joan Patterson and Hamilton McCubbin (1987)

as part of the Family Stress, Coping, and Health Project directed by Hamilton McCubbin.

An important characteristic of the A-COPE inventory is that the instrument is based on

theory that integrates individual coping theory and family stress theory (Moos & Billings,

1982). The authors worked from the perspective that healthy coping skills for adolescents

consists of successfully balancing the demands of the self, the family, and the community

(Patterson & McCubbin, 1987). Coping is viewed as synonymous with effective problem

solving, and this is achieved by maintaining a balance between utilizing both inner and

external resources to come to terms with difficulties (Patterson & McCubbin, 1987). The

items on the A-COPE were developed from literature review and from interviews with

adolescents regarding life changes.

The A-COPE (Patterson & McCubbin, 1987) is a 54-item questionnaire designed

to measure self-reported coping behaviors in adolescents between the ages of 13 and 18

years of age. Using a 5-point scale (never, hardly ever, sometimes, often, and most ofthe

time), participants are instructed to record how often they use each behavior by answering

the question, "When you face difficulties or feel tense, how often do you.. ?" The

A-COPE is scored by summing item scores for a total score; several items are reverse-

scored.








Each scale of the A-COPE assesses different coping patterns among adolescents.

For this study, the items from 6 of the 12 scales were included in the research

questionnaire. Those scales that were not necessary to achieve the purposes of this

research (e.g., cited unhealthy coping strategies such as using drugs) were excluded.

Reliabilities (cronbach's alpha) of the subscales that were used in this research are

high, ranging from .67 to .76 (Patterson & McCubbin, 1987). The reliabilities for each

scale used in this study are as follows: Developing self-reliance and optimism, alpha =

.69; Developing social support, alpha = .75; Solving family problems, alpha = .71;

Seeking spiritual support, alpha = .72; Investing in close friends, alpha = .76; and

Engaging in demanding activity, alpha = .67. Data on stability are not available.

However, reliability data from the Young Adult-COPE, which are only slightly modified

from A-COPE, show an overall alpha of .82 and good stability with a test-retest

correlation of .83 (Patterson, McCubbin, & Grochowski, 1983).

Concurrent validity was established initially by the authors by examining eight of

the coping scales in relationship to substance use among adolescents. Correlations for

female adolescents found two types of coping patterns: complementary coping patterns

that appeared to complement substance use, and competing coping patterns that appeared

to compete against substance use. For female adolescents, coping focused on solving

family problems (r = -. 10 to -. 21), seeking spiritual support (r = .11 to -. 21), and

engaging in demanding activity (r = .13 to -. 18), were negatively associated with the use

of cigarettes, beer, and marijuana (McCubbin, Needle, & Wilson, 1985).

The A-COPE has fair predictive validity, with correlations in predicted directions

with use of illicit substances including alcohol and marijuana. In a study by McCubbin,








Kapp, and Thompson (1993) of families of youth at risk involved in residential treatment,

youth coping was significantly related to program completion. Youth coping also related

to successful 3-month posttreatment adaptation. In a study by McCubbin et al.(1985),

adolescent coping was significantly related to adolescent health risk behaviors.

The original samples from which Patterson and McCubbin normed their

instrument lacked diversity in racial, cultural, and economic populations. Their three pilot

samples were each from the Midwest and were composed almost exclusively of

Caucasian adolescents from middle to upper socioeconomic families. Since then,

however, other researchers have used the instrument with more diverse groups of

adolescents and have reported that the instrument does appear to have validity with

diverse populations (Plancherel & Bolognini, 1995). For example, Copeland and Hess

(1995) report that they administered the A-COPE to 244 Anglo and Hispanic ninth-grade

students from a small, urban city. They found that the A-COPE differentiated coping

behaviors by both gender and ethnicity.

The Rosenberg Self-Esteem Scale (Rosenberg, 1965) is a scale that measures the

self-acceptance aspect of self-esteem. This measure was chosen because of the

instrument's focus on liking and approving of the self in general, as opposed to self-

evaluation in a specific context (e.g., social situations). The scale consists of 10

statements, such as, "On the whole, I am satisfied with myself," which reflect a positive

or negative self-evaluation.

The Self-Esteem Scale was originally designed to measure global self-esteem of

junior and senior high school adolescents (Rosenberg, 1965, 1989). The Self-Esteem

Scale taps the extent to which a person is generally satisfied with his or her life, considers








herself or himself worthy, holds a positive attitude toward him or herself, or alternatively

feels dissatisfied.

The Self-Esteem Scale benefits from a vast amount of research that has been

conducted using this scale with a wide range of groups over the years. Rosenberg's

original research (1965) on the Self-Esteem Scale was conducted with 5, 024 adolescents

of varying ethnic backgrounds from the state of New York. Subsequent research involved

thousands of college students, junior and senior high school students, and adults from a

range of professions and occupations.

Participants use a 4-point Likert scale ranging from strongly disagree (1) to

strongly agree (4) to indicate the extent to which each item is descriptive of them. Scores

for the scale are obtained by adding the participants' responses to the items, with higher

scores indicating higher levels of self-esteem. Positive and negative items are presented

alternately in order to reduce the danger of a respondent set.

The items were originally structured as a Guttman-type scale in which item

content moves sequentially from weaker to stronger expression of self-perceptions. The

Self-Esteem Scale has excellent internal consistency, as indicated by a Guttman scale

coefficient of reproducibility of.92 (Rosenberg, 1965). Two studies of two-week test-

retest reliability show correlations of .85 and .88, indicating excellent stability (Silber &

Tippett, 1965).

A great deal of research demonstrates the concurrent, known-groups, predictive,

and construct validity of the SES. The scale correlates 0.59 with Coopersmith's Self-

Esteem Inventory (Crandall, 1973), and from 0.56 to 0.83 with several similar measures

such as the Health Self-Image Questionnaire, providing evidence of convergent validity








(Robinson & Shaver, 1973; Silber & Tippet, 1965). Similarly, the Self-Esteem Scale

correlates in predicted directions with measures of depression, anxiety, and peer-group

reputation, demonstrating good construct validity by correlating with measures with

which it should theoretically correlate and not correlating with those with which it should

not. Construct-related validity has been demonstrated by numerous studies showing

correlations in appropriate direction between the Self-Esteem Scale scores and several

other variables (e.g., depression, anxiety) with which self-esteem theoretically is expected

to relate (Fischer & Corcoran, 1994).

Although some criticism has been aimed at the scale, primarily because of the

small number of items, enough research has accumulated to indicate that the measure

continues to be a powerful predictor of self-esteem in adolescents (Nunnally, 1967).

Dependent Variables

Indices of academic achievement consisted of two measures: (a) the differences

between fifth and eighth grade classroom grades in the subjects of math and reading, and

(b) the differences between fifth and eighth grade standardized test scores. Although

academic grades are sometimes subject to criticism due to the subjective manner in which

they can be assigned, their universality as a standard of performance seems to justify their

inclusion as a gauge of academic achievement in this study.

Data Analysis

A path analysis with manifest variables was used to test the hypothesis and

Research Question One and Research Question Two. Although the preferred analysis for

this study would have been path analysis with latent variables, it was not possible to use

this procedure because (a) the measures of the constructs of interest that were used in this








research contain more items than permissible, (b) the sample size was not sufficiently

large, and (c) it was not possible at the onset to predict how the variables would relate to

each other. All of these reasons combined to argue against using a path analysis with

latent variables (Hatcher, 1994).

Path analysis with manifest variables was a good choice for this study, although

admittedly not a perfect one. Path analysis with manifest variables assumes that each

assessment instrument perfectly measures the construct being measured, and although

each measure selected for this study is likely to be a good fit, it will not be perfect.

Despite this problem, path analysis with manifest variables was the best option for this

research because the use of path analysis with latent variables would be ill advised under

the current study's conditions, and because the use of simple correlations would not

assess the complex relationships hypothesized among the variables in this study.

Path analysis is a form of multiple regression that uses regression correlations in a

visual diagram to find correlation paths among the relevant variables. The use of path

analysis to analyze this research data allowed the portrayal of a theoretical picture

indicating which paths between and among variables proved to be statistically significant.

A theoretical picture was portrayed at the onset of the research that indicated the

hypothesis and research questions. After data analysis was complete, the nonsignificant

paths were eliminated from the theoretical picture, and a path diagram was generated

indicating the results of the data collected in this study (Hatcher, 1994). Figure 1

illustrates the initial theoretical model for this research.








Hypothesis and Research Questions

Hypothesis

As stereotypical gender role beliefs among early adolescent females increase, and

as healthy coping strategies and self-esteem decrease, it was predicted that academic

achievement scores from fifth to eighth grade would decrease. A path analysis with

manifest variables was used to test whether the predictor variables of stereotypical gender

role beliefs, coping strategies, and self-esteem influence the two criterion variables of

fifth and eighth grade standardized test score differences and fifth and eighth grade math

and reading classroom grade differences.

Research Question 1

How do the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem correlate with each other? Three Pearson product-moment

correlation coefficients were calculated to assess the magnitude of the correlation

between pairs of these three predictor variables.

Research Question 2

Which pairs of the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem have the greatest impact on fifth and eighth grade math and

reading classroom grades differences and fifth and eighth grade standardized test score

differences? A path analysis with manifest variables was used to test the six indirect

correlational paths between the predictor variables and the criterion variables.

Research Question 3

What is the relationship between the two criterion variables of fifth and eighth

grade standardized test score differences and fifth and eighth grade math and reading




















































Figure 1. Conceptual model of the predictors of academic achievement.










classroom grade differences? Correlation coefficients were used to test the correlations

between the two criterion variables.

Research Question 4

Will the relationships between stereotypical gender role socialization, coping

skills, self-esteem and academic achievement for middle school females vary by ethnicity

and socioeconomic status? These data were analyzed by the use of Pearson product-

moment correlations and z tests. By obtaining the Pearson product-moment correlation

coefficients and performing r to z transformations, I was able to perform z tests. Forty-

two pairs of correlations were compared using z tests, with the Benferroni-corrected

alpha level of p < .0025 (.05/20 = .0025). The Forty-two pairs of correlations included 21

pairs for race and 21 pairs for socioeconomic status.

Research Question 5

Will the mean scores of stereotypical gender role socialization, coping skills, self-

esteem, and academic achievement for middle school females vary by ethnicity and

socioeconomic status? These data were analyzed by 14 independent t tests, 7 for race and

7 for socioeconomic status. The independent variables are race or socioeconomic status,

and the dependent variables were stereotypical gender role socialization, coping skills,

self-esteem, fifth and eighth grade standardized test score differences and fifth and eighth

grade classroom grade differences.

Methodological Limitations

One possible limitation of this research was the use of the subscales of the coping

measure employed in this research. The A-COPE was originally designed as a 12-scale








instrument. Six of the subscales represent positive coping skills, and because the use of

these six fulfilled the measurement needs for this research, only these six scales were

utilized in this study. It should be noted that other researchers, including the instrument's

authors, have used partial scales from the A-COPE and found positive results (Kluwin,

Blennerhassett, & Sweet, 1990; McCubbin et al, 1985; Needle, Su, Doherty, Lavee, &

Brown, 1988). The reliability coefficients for each subscale of the A-COPE are listed in

the instrumentation section of this chapter.

In addition to these reasons for using six of the subscales of the A-COPE, it is

noteworthy to consider the effect of context in this study, or question context effect.

Participants in this study are young adolescents in a school setting who will be answering

questions for an adult who is a stranger to them. Given the school context in which they

will be responding to this measure, it is likely that many students may not be entirely

truthful if asked to answer questions about drug use or swearing, for example. The six

subscales of the A-COPE that were not utilized for this research contain items that

inquire about the respondent's use of more negative coping behaviors; thus it is possible

that the introduction of negativity could lead to distortion of response on the positive

items as well.

A second potential limitation of this study is the timing of the data collection. The

population used in this study was composed of ninth grade students who completed the

three assessment measures early in their academic year. However, the academic

achievement measures are from the end of their fifth and eighth grade school years.

Therefore, there was a difference of approximately six months' time that elapsed between








the two sets of measurements. This difference in timing needs to be taken into account

when evaluating the findings of this study.

Finally, the criteria used in this study for socioeconomic level is the students'

qualification for free or reduced lunches according to the standards set by the State of

Florida's Department of Education. This measure of socioeconomic status was chosen

over inquiries about parents' highest level of educational attainment because the lunch

program seems to be a more direct assessment of the desired information. In addition,

permission to ask the students about their parents' educational attainment was frowned

upon by officials from several of the participating school boards. Limitations do remain

regarding the lunch program, however. First, the status of qualification for the free or

reduced lunch program simply divides students into two categories: below the cut-off line

or above the cut-off line. Thus, respondents were categorized as simply lower

socioeconomic status or higher socioeconomic status. Secondly, there is a general under-

utilization of the lunch program in the county as students progress in public schools, i.e.,

the same family with the same income requests free or reduced lunches in elementary

school but may not in high school. Therefore, although this measure is believed to be as

accurate a gauge of socioeconomic level that can be obtained for this sample, the actual

socioeconomic status of the sample may not be fully represented by this measure














CHAPTER IV
RESULTS OF THE STUDY

The purpose of this study was to examine the relationship between adolescent

females' academic achievement in middle school and traits of resilience, including high

self-esteem, healthy coping skills, and low adherence to stereotypical gender role beliefs.

Additional research questions included an investigation each of these variables

individually and in pairs by race and class. This study attempted to answer the following

questions:

1. Do gender role beliefs, coping strategies or self-esteem predict academic

achievement for middle school adolescent females?

2. How do gender role beliefs, coping strategies, and self-esteem correlate with

each other for middle school females?

3. When gender role beliefs, coping strategies, and self-esteem are paired

together, which has the greatest impact on academic achievement for females over

middle school?

4. What is the relationship between changes in standardized test scores and

changes in classroom grades for females over the period of middle school?

5. Will the relationships among gender role beliefs, coping skills, self-esteem and

academic achievement for middle school females vary by race or socioeconomic status?








6. Will the mean scores of gender role beliefs, coping skills, self-esteem, and the

changes in academic achievement over middle school vary for females by race or

socioeconomic status?

Descriptive Data

A total of 591 students participated in this research, although data from just 291

were used in this study. The assessment measurement was administered to 591 male and

female ninth-graders in classrooms of eight different schools. Five public school districts,

representing Alachua, Bradford, Clay, and Putnam counties and the University of Florida

research laboratory school are represented in the sample.

As follow up data collection ensued with the various school administrators, it

quickly became apparent that a sizable number of student records were incomplete, thus

making it impossible to locate the matching academic records necessary for this research.

Therefore, although 357 girls completed the research survey, full information for just 291

of the students was available at the time of data collection, and it was with these data that

this study was completed.

Several different standardized tests are represented in the research results.

Students who transferred to the five districts from other schools, counties, or states

sometimes had differing standardized tests scores from their previous school districts.

Therefore, although the majority of the research participants had standardized test records

from either the Iowa Tests of Basic Skills (ITBS), or the Comprehensive Tests of Basic

Skills (CTBS), other tests are also represented in this sample, including the California

Achievement Tests, the Stanford Achievement Test and the Florida Comprehensive

Achievement Test.








In some cases student records indicated that standardized tests were taken at the

end of fourth grade, although most frequently the tests were administered at the end of

fifth grade. The data included in this research, therefore, include records from the

students' last recorded elementary school standardized test score, whether that was fourth

or fifth grade.

Each standardized test produces several scores in each subject area tested. For

example, the ITBS reports three reading scores: vocabulary, reading and total reading,

and four math scores: math concepts, math problems, math computation, and math total.

For this research the "total reading" and "total math" scores were recorded as reported in

the form of students' national percentile ranking.

A few participants attended elementary schools that did not assign letter grades

for subjects but used another comparable system such as citing "mastery of skills,"

"partial mastery of skill," or "needs improvement." In these cases, these descriptive

evaluations were assigned a number score, with "4" equivalent to an "A" and "1"

equivalent to a "D" and then tallied to derive a letter score consistent with the letter

grading format reported by the majority of student participants.

There were some differences that emerged as a result of citing course subjects

from different school districts. For example, although most schools listed "reading,"

some schools listed "language arts" and some schools listed both as separate subjects.

Most eighth graders took a course called "math" but some did not take math but took

"algebra" instead. For this data collection, the subjects of "reading" and "language arts"

are used interchangeably. In the case of a student having two reported classes in the

subject of math or reading, both grades were noted and an average grade was derived.








As discussed in Chapter III, in order to analyze the data by socioeconomic class

and race, a minimum of 30 respondents for each category was needed. Data collection

continued until enough data were collected to permit analyzing findings for White and

African American females, and for low- and middle- or high-income students. The total

sample consisted of 75 African American females (25.8%), 206 White females (70.8%),

8 Hispanic females (2.7%), 1 Asian American female (0.3%), and I Multiethnic female

(0.3%). The sample included 106 low income students (36.4%) as reported by the school

district as qualified to receive free or reduced lunch, and 185 students (63.6%) of higher

income families, as reported by the school district as not qualified to received free or

reduced lunch. The ages of the ninth-grade students ranged from 13 years to 16 years of

age, while the average age of the research participant was 14 years and 4 months of age.

Student participants for this study included all of the different ability levels

represented at the eight public schools. However the level of consistency of reporting of

these levels is poor, i.e., some schools reported their low ability level students and some

did not, so therefore the breakdown of this demographic factor is not included in this

report. What is known is that this research sample consists primarily of students of

normal ability as well as some students from varying exceptionality categories which

typically include students with learning, behavioral, and emotional problems, as well as

student who qualify for gifted or "honors" classes. There is no reason to think that the

ability levels represented in this sample are not a normal representation of typical ninth

grade student bodies.








Hypothesis

As stereotypical gender role beliefs among early adolescent females increase, and

as healthy coping strategies and self-esteem decrease, it was predicted that academic

achievement scores from fifth to eighth grade would decrease. A path analysis with

manifest variables was used to test whether the predictor variables of stereotypical gender

role beliefs, coping strategies, and self-esteem would influence the two criterion variables

of fifth and eighth grade test score differences and fifth and eighth classroom grade

differences.

Students with the highest self-esteem demonstrated the least decline in math

grades from fifth to eighth grades, substantiating the hypothesis on one indicator by the

results of this research. Data analysis indicated that the predictor variable of self-esteem

significantly predicted the differences between fifth and eighth grade classroom grades in

math, r = .15, p < .02. Females with the lowest self-esteem demonstrated the greatest

decrease in math grades over middle school. Figure 2 presents a diagram of the results of

the path analyses for the hypothesis that yielded statistically significant results. The path

coefficients shown in Figure 2 were obtained from the results of four simultaneous, linear

multiple regression analyses, one for each of the four criterion variables: (a) change in

math grades, (b) change in reading grades, (c) change in math standardized test scores,

and (d) change in reading standardized test scores. The predictor variables were scores on

the A-COPE, Rosen, and PAQ, and the fifth grade classroom grade or standardized test

scores that corresponded to the particular criterion variable (e.g., fifth grade math

classroom grade for the regression with math grade change).








In contrast, all other analyses failed to show a significant relationship between the

predictor and criterion variables. Specifically, the predictor variables of stereotypical

gender role beliefs and coping skills did not significantly predict differences between

either fifth and eighth grade classroom grades in reading or math, nor the differences

between fifth and eighth grade standardized test score differences in reading. The

predictor variable of self-esteem did not significantly predict the differences between

fifth and eighth grade classroom grades in reading, nor the differences between fifth and

eighth grade standardized test scores in reading or math. Figure 2 presents a diagram of

all significant and nonsignificant path analyses for the hypothesis. Figure 3 presents a

diagram of the results of all significant path analyses for the hypothesis. Table 1 presents

the simple bi-variate correlations for all variables employed in the path analysis for this

study.

Research Question 1

How do the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem correlate with each other? Three Pearson product-moment

correlation coefficients were calculated to assess the magnitude of the correlation

between pairs of these three predictor variables.

Young women who scored the highest on the PAQ also scored the highest on the

A-COPE, and young women who scored the highest on the Rosenberg Self-Esteem Scale

also scored the highest on the A-COPE, thus substantiating two of the three predicted

correlations of Research Question 1. Specifically, stereotypical gender role beliefs

correlated significantly with healthy coping strategies, r = .44, p < .0001 and r2 = 19%,










































Note. R = reading. M = math. For correlations among the predictor variables, n = 291.
For correlations among the predictor variables and the criterion variables for reading
grades n = 266, for math grades n = 263, for reading standardized test scores n = 261, and
for math standardized test scores n = 258. For correlations between the criterion variables
for reading n = 206 and for math n = 203.
*p < .05. **p <.005. *** p <.0001.



Figure 2. Significant and nonsignificant path analysis results of predictors of
academic achievement, the relationships among predictors, and the relationship between
grades and standardized tests.

















Academic
achievement

Grades


Academic
achievement

Standardized Tests


Note. M = math. For the correlations among the predictor variables, n = 291. For the
correlations between Rosen and math grades n = 203, and for the correlations between
the criterion variables n = 203.
* p <.05. **p <.005. *** p <.0001.



Figure 3. Significant path analysis results of predictors of academic achievement, the
relationships among predictors, and the relationship between grades and standardized
tests.









Table 9
Simple Bi-Variate Correlations for Predictor and Criterion Variables


Variable 1 2 3 4 5 6 7 8 9 10 11


1. PAQF 1.0

2. ROSEN .12* 1.0

3. A-COPE .44* .39* 1.0

4. RDG 5 .14* .12 .01 1.0

5. RDG 8 .10 .15* .02 .41* 1.0

6. MTH G 5 .17* .07 .06 .74* .41* 1.0

7. MTH G 8 .06 .11 -.04 .33* .59* .42* 1.0

8. RDST 5* .05 -.07 -.09 .49* .41* .47* .34* 1.0

9. RDST8 -.04 -.07 -.15* .44* .36* .42* .36 .80* 1.0

10. MTH ST5 .02 .02 -.12 .52* .38* .58* .51* .65* .61* 1.0

11. MTH ST 8 -.00 .02 -.11 .51* .43* .55* .52* .61* .68* .75* 1.0


Note. p < .05. n ranges from 220 to 291.








indicating that 19% of the variance of PAQ scores overlapped with A-COPE scores.

Secondly, high self-esteem correlated significantly with healthy coping strategies, r = .39,

p < .05 and r2 = 15%, indicating that 15% of the variance of the Rosenberg scores

overlapped with A-COPE scores.

In contrast, stereotypical gender role beliefs were found to have a low correlation

with high self-esteem, r = .12, p < .05 and r2 = 10%. Figure 2 presents a diagram of all

significant and nonsignificant paths for Research Question 1. Figure 3 presents a diagram

of the results of only the significant paths only Research Question 1.

Research Question 2

Which pairs of the predictor variables of stereotypical gender role beliefs, coping

strategies, and self-esteem have the greatest impact on the criterion variables of fifth and

eighth grade test score differences and fifth and eighth grade classroom grade

differences? A path analysis with manifest variables was used to test the six indirect

correlational paths between the predictor variables and the criterion variables of fifth and

eighth grade test score differences and fifth and eighth grade classroom grade differences.

The results of Research Question 2 indicated one significant finding. The interaction of

stereotypical gender role beliefs and coping strategies significantly predicted changes in

math classroom grades, at p > .01. The regression coefficients used to assess Research

Question 2 are shown in Table 2 and were obtained from the results of four, simultaneous

linear multiple regression analyses, one for each of the criterion variables (a) change in

math grades, (b) change in reading grades, (c) change in math standardized test scores,

and (d) change in reading standardized test scores. The predictor variables were the three

two-way interactions of scores on the A-COPE, Rosen, and PAQ, and the fifth grade








classroom grade or standardized test score that corresponded with the particular criterion

variable (e.g., fifth grade reading classroom grade for the regression with reading grade

change).

In contrast, none of the other interactions between stereotypical gender role

beliefs, coping strategies, and self-esteem significantly reached or approached

conventional levels of statistical significance to predict changes in either standardized test

scores in reading or math, or changes in classroom grades in reading or math. The

significant data results in this research question are noteworthy and yet, upon further

consideration, it was recognized that the analysis as stated did not indicate the full

implications of this finding. Data analysis from the hypothesis revealed that individually

neither adherence to stereotypical gender role beliefs nor coping skills predicted

academic changes, yet analysis from this research question indicates that the interaction

between these two variables was a significant predictor of math grades. The question

remained: in what direction were the significant changes in math grades? Who scored the

highest in math standardized tests?

Further analysis of the significant findings of Research Question 2 was pursued

via the post-hoc analysis of Least Squares Means set at p < .05. Two significant results of

the least square mean effect were found. First it was found that the least square mean

value of- 0.16 for PAQ1 and A-COPE 1 was significantly different than the least square

mean value of 0.76 for PAQ3 and A-COPE 2 at p <.05. Secondly, it was found that the

least square mean value of- 0.13 for PAQ1 and A-COPE 2 was significantly different

than the least square mean value of 0. 76 for PAQ3 and A-COPE 2, at p <.05.








The post-hoc analysis revealed that participants who scored low in stereotypical

gender role beliefs and low or medium in coping showed the least decline in math grades.

These results indicate that for this sample low PAQ and low A-COPE predicted a math

grade decline of 15/100 of a grade point in contrast to the combination of high PAQ and

medium A-COPE which predicted a steeper decline of 76/100 of a grade point. Research

participants who scored high in stereotypical gender role beliefs and moderate in coping

skills showed the most significant loss in math classroom grades, by 76/100 of a grade, in

contrast to participants who were low in stereotypical gender role beliefs and low or

medium in coping who showed a fairly small decline in math grade, 13/100 15/100 of a

grade.

Research Question 3

What is the relationship between the change in classroom grades and the change

in standardized test scores in math and reading? Correlation coefficients were used to test

the correlations between the two criterion variables of the change between fifth and

eighth grade classroom grades and the change between fifth and eighth grade

standardized test scores.

Declines in reading grades from fifth to eighth grades do not have corresponding

declines in standardized test scores, according to the results of Research Question 3. Data

analysis revealed there was a significant relationship between the change in fifth and

eighth grade grades and the change in fifth and eighth grade standardized test scores in

math, r = .26, p < .005. In contrast, analysis revealed that there was no significant that

relationship between changes in fifth and eighth classroom grades and fifth and eighth

grade standardized test scores, r = .15, p < .15. In other words, changes in math grades








Table 2

Summary of Simultaneous Regression Analyses for the Interaction of Pairs of Variables
Predicting Changes in Four Measures of Academic Achievement


Predictor Variables B SE B


Reading Grade

1. PAQF x A-COPE -.03 .03 -.10
2. PAQF x ROSEN .00 .00 .17
3. A-COPE x ROSEN -.00 .00 .00



Math Grades

1. PAQF x A-COPE -.07 .03 -.21
2. PAQF x ROSEN .01 .00 .21
3. A-COPE x ROSEN .00 .00 .02



Reading Test Scores

1. PAQF x A-COPE -.46 .45 -.06
2. PAQF x ROSEN .00 .05 .00
3. A-COPE x ROSEN -.03 .05 -.05



Math Test Scores

1. PAQF x A-COPE -.06 .55 .00*
2. PAQF x ROSEN .02 .06 .03
3. A-COPE x ROSEN -.01 .06 -.01


Note. *p < .05.









from fifth to eighth grades have corresponding changes in standardized test scores, but

changes in reading grades from fifth to eighth grades do not have corresponding changes

in standardized test scores. Figure 2 presents a diagram of all significant and

nonsignificant path analyses for Research Question 3. Figure 3 presents a diagram of the

results of the significant path analyses only for Research Question 3.

Research Question 4

Will the relationships among stereotypical gender role socialization, coping skills,

self-esteem and academic achievement for middle school females vary by race or

socioeconomic status? The data were analyzed by the use of Pearson product-moment

correlations and z tests. By obtaining the Pearson product-moment correlation

coefficients and performing rto z transformations, I performed 42 z tests. The 42 pairs of

correlations included 21 pairs for race and 21 pairs for socioeconomic status. The

Bonferroni-corrected alpha level was p < .21 (.05/21 = .0025).

All results for this research question were nonsignificant and an analysis of the

totals was also nonsignificant. First, results indicated that the average correlations were

not significantly different for research participants by race, indicating that the

relationships between stereotypical gender role socialization, coping skills, self-esteem

and academic achievement for middle school females in this sample did not vary by race.

Second, results indicated that the average correlations were not significantly

different for research participants by socioeconomic status, indicating that the

relationships between stereotypical gender role socialization, coping skills, self-esteem

and academic achievement for middle school females did not vary by socioeconomic

status in this research sample.




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EE5YPX6N7_UQLDYG INGEST_TIME 2013-10-24T16:52:00Z PACKAGE AA00017712_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES