Group Title: development and validation of a multidimensional exploration and commitment scale for assessing ego identity development
Title: The development and validation of a multidimensional exploration and commitment scale for assessing ego identity development
CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102788/00001
 Material Information
Title: The development and validation of a multidimensional exploration and commitment scale for assessing ego identity development
Physical Description: Book
Language: English
Creator: Moberly, Jeanne C., 1943-
Copyright Date: 1985
 Record Information
Bibliographic ID: UF00102788
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: ltuf - AEG4431
oclc - 14368020

Full Text















THE DEVELOPMENT AND VALIDATION OF A MULTIDIMENSIONAL
EXPLORATION AND COMMITMENT SCALE FOR
ASSESSING EGO IDENTITY DEVELOPMENT








By

JEANNE C. MOBERLY


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


1985
























Copyright 1985

by

Jeanne C. Moberly























Dedicated to my two children, Laura and Richard,
whose pride in me has been a constant source of
support throughout this venture, and of whom I am
very proud.















ACKNOWLEDGEMENTS


I would like to thank my committee chairperson,

Dr. Dorothy Nevill, who has been a good friend as well

as a mentor throughout my graduate school experience.

I would also like to thank my committee members, Drs. Grater,

Resnick, Miller, and Fukuyama, whose suggestions and ques-

tions helped me to clarify and articulate my own ideas.

There are many friends I wish to acknowledge--people who

xeroxed and assembled tests and aided me in collecting

and coding the data: Dr. Robert Ashley, Pat Atkins, Barbara

Beynon, Betty Black, Cheryl Boggess, Drs. Leonard Beeghley

and Mary Anna Hovey, Lissa Friedman, Tricia Gregory, Dale

Midgett, Kristen McIntyre, Marilyn O'Connor, Sandy Peterson,

Joyce Perrotta, and Ann Wilson. Thank you for your time

and efforts, energy and support.















TABLE OF CONTENTS

Page

ACKNOWLEDGEMENTS....... ............................. iv

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

CHAPTERS

I INTRODUCTION.............................. 1

II REVIEW OF THE LITERATURE.................. 8

Erikson's Psychosocial Stages............. 8
Adolescence ................... ........... 10
Definition of Ego Identity................ 11
Ego's Tasks in Identity Development....... 12
Seven Perspectives on Ego Identity........ 13
Early Attempts to Measure Differences in
Ego Identity Development................ 16
Marcia Interview. ......................... 20
Other Measures of Ego Identity............ 24
Gender Difference in Ego Identity
Development............................. 26
Separating Content Areas.................. 36
Exploration and Commitment Variables...... 39
Social Desirability........................ 41
Instrument Development--Reliability and
Validity ................................. 41
Hypotheses... ............................. 46

III METHOD... ................................. 55

Subjects. ................................ 55
Procedure ................................ 56
Instruments................................ 62

IV RESULTS ................................... 68

Reliability ................................. 69
Content Validity.......................... 70
Construct Validity........................ 73
Concurrent Validity....................... 76
Social Desirability Bias .................. 85
Summary.................................... 88








V DISCUSSION... ... ......................... 89

Reliability Data.......................... 89
Content Validity.......................... 89
Construct Validity........................ 91
Concurrent Validity....................... 94
Social Desirability ....................... 106
Summary................................... 107
Future Directions......................... 109


APPENDICES

A SELF-REPORT MEASURE OF EXPLORATION AND
COMMITMENT.............................. Ill

B EXPLORATION AND COMMITMENT SCALE......... 114

C CROWNE-MARLOWE SOCIAL DESIRABILITY SCALE.. 124

D SIMMONS IDENTITY ACHIEVEMENT SCALE........ 127

REFERENCES............................................. 130

BIOGRAPHICAL SKETCH............................. 139














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

THE DEVELOPMENT AND VALIDATION OF A MULTIDIMENSIONAL
EXPLORATION AND COMMITMENT SCALE FOR ASSESSNG
EGO IDENTITY DEVELOPMENT

By

Jeanne C. Moberly

May 1985

Chairperson: Dorothy Nevill
Major Department: Psychology

The major goal of this research was to develop a reliable

and valid instrument for assessing ego identity development.

The two variables that have consistently been regarded as

integral parts of ego identity are (1) the individual's

experience of exploration of his or her options and values

and (2) the development of a sense of commitment to values

that provide self-definition.

The Exploration and Commitment Scale (ECS) is a 64-item

instrument containing 32 exploration items and 32 commitment

items, each scored on a 7-point scale. The ECS is divided

into four subscales (Career, Interpersonal, Sexual, and

Ideology), each containing eight exploration and eight com-

mitment items. The ECS may be scored to provide the following

information:

1. an overall score--a measure of ego identity
development,


vii









2. an overall exploration score,

3. an overall commitment score,

4. an Identity Status (Achievement, Moratorium,
Foreclosure, or Commitment) within each subscale,

5. an exploration and/or commitment score for each
subscale.

Data analyses from 245 undergraduates at the University

of Florida indicate a moderately high level of internal

consistency and a significant level of test-retest reliabil-

ity for the ECS. The convergence and divergence of subscale

items supports the belief that the subscales measure dis-

criminably different areas of identity. Correlations with

a self-report measure indicate that test results compare

significantly with the individual's own experience of

exploration and commitment. Correlations with an instrument

that measures overall ego identity development were also

significant. Correlation with the Crowne-Marlowe Social

Desirability Scale confirmed a lack of social desirability

bias in the responses. Comparisons between males and females

within the subscales indicate that exploration and commitment

within the Interpersonal and Sexual areas are particularly

salient for females. Males have explored more and express

more commitment in the Ideology area. Small between-sex

differences in the Career Commitment area indicate an equally

high level of career commitment for both sexes.

The validity and reliability of the ECS were supported.

Directions for future utilization of the measure were

suggested.


viii














CHAPTER I
INTRODUCTION


Erik Erikson (1959, 1968) is generally considered

to be the most influential theorist and writer on the

processes of ego identity development. He placed the crucial

period for ego identity formation in adolescence and

described both the psychodynamic bases of and the psycho-

social influences on ego identity.

Operationalizing the concept of ego identity is the

first step in studying its processes. A variety of defini-

tions can be seen in the early attempts to construct measures

of ego identity (Bronson, 1959; Block, 1961; Heilbrun,

1964; Gruen, 1960; Hauser, 1972; Constantinople, 1969;

and Dignan, 1965.

Other measures of ego identity have been constructed

so that they are closely linked conceptually to Erikson's

theories (Rasmussen, 1964; Marcia, 1964, 1966; Baker, 1971;

Simmons, 1970, 1973; Tan, Kendis, Fine, and Porac, 1977;

Schilling, 1975; and Adams, Shea, and Fitch, 1979). The

semistructured interview developed by James Marcia (1964,

1966) has been the most widely used method for operationaliz-

ing adolescent ego identity development (Bourne, 1978a).

The interview format categorizes individuals into identity

statuses which Marcia describes as "four modes of dealing

1









with the identity issue characteristic of late adolescents"

(1980, p. 161). Marcia's statuses (Achievement, Moratorium,

Foreclosure, and Diffusion) are closely related to Erikson's

formulation of ego identity development.


There are two notions which seem to charac-
terize Erikson's writing about the phase
of late adolescence. The first is the
presence of some period of re-thinking,
sorting through, trying out various roles
and life plans. This is taken to be the
behavioral referent for the internal process
referred to in psychoanalytic theory as
"The synthesizing function of the ego"
(Erikson, 1956, p. 104). We call this
decision period, crisis, a time during
adolescence when the individual seems to
be actively involved in choosing among
meaningful alternatives.
The second indication of the achievement
of ego identity is subsequent commitment,
particularly in such important life areas
as occupation and ideology. Commitment
refers to the degree of personal investment
the individual expresses in a course of
action or belief. (Marcia, 1964, pp. 23-24)


Marcia's statuses are defined based on 1) the presence

or absence of a period of exploration of alternatives and

decision making ("crisis period") and on 2) the degree

of commitment exhibited in areas considered crucial to

ego identity formation. An individual in the achievement

status has gone through a period of exploration and has

developed firm commitments to self-chosen goals. Moratorium

youths are in the midst of a crisis period of exploration

and are moving toward firm commitments, though they have

not yet clarified their goals. Foreclosures exhibit strong









commitments, but their commitments (usually parentally

influenced) have been arrived at without having experienced

a period of exploration. Youths who are assigned to the

diffusion status demonstrate no clear or firm commitments,

nor do they appear to be currently exploring to clarify

their values and goals. Matteson (1977) underlined the

importance of the variables of exploration and commitment

to ego identity development in his study with Danish adoles-

cents. He suggested measuring the variables separately

and on continuous scales.

The interview content areas explored in research on

ego identity have varied across studies. Marcia's original

content areas for use with males were career and political

and religious ideology. When the interview was extended

to females, "attitudes toward sexual intercourse" was added

to the interview (Marcia and Friedman, 1970). Sex role

ideology and sex values have also been utilized as content

areas of the interview (Matteson, 1977; Hopkins, 1980;

Nevid, Nevid, O'Neill, and Waterman, 1974). Interpersonal

areas (friendship and dating) have recently been suggested

as additional areas to consider in identity formation

(Grotevant, Thorbecke, and Meyer, 1982; Thorbecke and

Grotevant, 1982). As the number of interview areas has

increased, an overall identity status has become less and

less meaningful.









The importance of each of the content areas to ego

identity development has been explored for both males and

females (Matteson, 1977; Nevid et al., 1974; Schilling,

1975; Hodgson and Fischer, 1979; Waterman and Nevid, 1977;

Kacergius and Adams, 1980; Thorbecke and Grotevant, 1982;

Waterman and Waterman, 1971; Waterman, Geary, and Waterman,

1974; and Rogow, Marcia, and Slugowski, 1982). It has

become clear that an increased understanding of ego identity

development will be assisted by separate evaluation of

the content areas. Sex differences in the importance of

the content areas to identity development is an especially

important area of investigation.

The Marcia Interview used in much of the above research

has been criticized as being too time-consuming. Since

it takes between 30 to 45 minutes to administer, sample

sizes in studies using the interview are typically small.

Additionally, the ratings used to categorize individuals

into statuses rely on subjective judgment (Schilling, 1975;

Adams, Shea, and Fitch, 1979; Simmons, 1970).

Schilling (1975) and Adams, Shea, and Fitch (1979)

have both created questionnaires designed to replace the

Marcia Interview in assigning subjects to an overall identity

status. Because they can be group administered, they are

less time-consuming, and the sample sizes can be increased.

Since the scoring procedures are clear and objective, they









avoid the problems of subjectivity associated with trained

raters assessing interviews to assign statuses.

The Schilling (1975) measure allows for separate assign-

ment of individuals to identity statuses based on either

the Interpersonal area or the Occupational, Political,

and Religious area. Schilling provided little evidence

of content or construct validity, internal consistency,

or test-retest reliability in presenting his measure which

was written for use with males.

Adams, Shea, and Fitch (1979) present considerable

evidence of internal consistency, predictive validity,

construct validity, and test-retest reliability with their

measure which has been used with large numbers of male

and female college undergraduates. The measure, as it

is presently constructed, has some limitations. It provides

only for assignment to an overall identity status based

on the three original content areas used with males--

occupation and political and religious ideology. It does

not provide for a separate assessment of the exploration

and commitment variables. And, scoring procedures allow

for "transition statuses" (i.e. Achievement-Diffusion or

Foreclosure-Moratorium) that are theoretically inconsistent.

There is still a need for an objective measure of

ego identity that assigns individuals to an identity status

in each content area of identity--Interpersonal (friendship

and dating), Sexual Values and Sex Roles, Career, Political









and Religious Ideology--based on test items that clearly

reflect the presence or absence of exploration and commitment

variables, and that allow for separate analyses of these

two variables. The purpose of this study is to create

a reliable and valid objective measure of ego identity

that has the following qualities:

1. It includes a variety of areas that have been

shown to contribute to ego identity. These areas include

the Interpersonal (Dating and Friendship), Sexual and Sex

Role Values, Career, and Political and Religious Ideology.

2. It is based on the amount of exploration and the

level of commitment exhibited in these areas and allows

for separate assessment of each of these variables.

3. It allows for the assignment of a separate ego

identity status for each content area.

4. It enables the calculation of a total score which

reflects the current level of identity achievement in terms

of the amount of exploration and level of commitment the

individual has experienced.

5. It is easily and quickly administered and scored.

6. Scoring is not influenced by the subjectivity

of raters, and responses are not influenced by social

desirability concerns.

The following literature review covers theoretical

conceptions of ego identity and the efforts made at opera-

tionalizing the concept. Research in the measurement of







7

gender differences in ego identity development is also

discussed. The content areas that have been used in assess-

ing ego identity are presented, and a discussion of the

research exploring the relative importance of the various

content areas for males and females is also included.














CHAPTER II
REVIEW OF THE LITERATURE


Erikson's Psychosocial Stages


Erikson (1959) defined psychosocial stages of develop-

ment from birth through old age (Chart 1). Each stage

occurs within a particular age period. The early stages

parallel Freud's psychosexual stages, but Erikson's stages

continue into old age. Each stage has a developmental

crisis associated with it. Each crisis grows out of the

individual's developmental readiness and society's demands

of the individual. The crisis arises at a particular time

developmentally but is not confined to one developmental

period. How an individual has mastered past crises will

influence how he (or she) masters later crises. The indi-

vidual's ability to meet each crisis adequately develops

within him or her the conviction that his (her) way of

being is


a successful variant of the way others in
the society master experience and recognize
mastery. Thus, self-esteem, confirmed at
the end of each major crisis, grows to be
a conviction that one is learning effective
steps toward a tangible future, that one
is developing a defined personality within
a social reality which one understands.
(Erikson, 1959, p. 89)








A

Psychosocial
Crises


CHART I
B

Radius of Significant
Relations


C

Psychosocial
Modalities


D

Psychosexual
Stages


I Trust vs.
Mistrust

II Autonomy vs.
Shame, Doubt

III Initiative vs.
Guilt



IV Industry vs.
Inferiority


V Identity
Achievement vs.
Identity
Diffusion

VI Intimacy vs.
Isolation


VII Generativity
vs.
Self-Absorption

VIII Integrity vs.
Despair


Maternal Person


Parental Persons


Basic Family




"Neighborhood,"
School


Peer Groups and
Outgroups;
Models of
Leadership

Partners in friend-
ship, sex, competi-
tion, cooperation

Divided labor and
shared household


"Mankind"
"My Kind"


To get
To give in return

To hold (on)
To let (go)

To make
(=going after)
To "make like"
(=playing)

To make things
(=completing)
To make things togeth

To be oneself
(or not to be)
To share being
oneself

To lose and find
oneself in
another

To make be
To take care of


To be, through
having been
To face not being


Note. From "Identity
Issues, 166.


and the life cycle," by E. H. Erikson, 1959, Psychological


Oral-respiratory,
Sensory-Kineshetic

Anal-Urethral
Muscular

Infantile-Genital
Locomotor
(Intrusive,
Inclusive)

"Latency"

.er

Puberty




Genitality









Adolescence


Adolescence has been described by many as a period

of transition between childhood and adulthood. In the

mastering of the crises of the first four stages of develop-

ment, the individual is presumably adequately prepared

to integrate an identity out of the growth and change of

the adolescent period.

Adolescence is a fertile period of change. Physical

and cognitive growth occurs which promotes psychological

changes and affects the way adolescents interact within

their social environment. Puberty triggers many changes.

As body shape becomes more adult-like, adolescents struggle

with changing self-images. As adolescents' sexuality

matures, they move away from their parents toward increas-

ingly more intimate peer relationships. As the distance

between parent and child becomes clearer, the adolescent

is able to gain a greater appreciation for the unique

attributes that make him or her distinct from others.

Cognitively, individuals move from the concrete cogni-

tive level to the level of formal operations. This movement

is characterized by an increasing ability to think abstractly

and to synthesize information. The world around the adoles-

cent can now be considered from a new, more wholistic,

and more integrated perspective. The adolescent is able

to confront abstract values and moral issues. The idealism

and questioning of human values, which is a familiar part









of adolescence, encourages growth in moral development.

The ability to empathize with another, which is a part

of a more mature level of moral development, is an important

foundation for the development of intimate relationships,

the next major life task.

The relationship between the developing individuals

and society is reciprocal. Social expectations influence

identity formation. The adolescents' self-definition must

be confirmed by the social world around them. Adolescence

has its beginnings in the physical changes of puberty and

its endings in the assumption of adult roles and responsi-

bilities through marriage, procreation, and self-support.

It is a time of preparation for these future adult roles.

Josselson (1980) emphasizes that developing ego identity

is a two-fold task:

1. Constructing a stable, integrated self, autonomous

from internal and external authority (Who am I?).

2. Fitting the new found self to the world (What

shall I become?).


Definition of Ego Identity


James Marcia (1980) has provided a concise definition

of ego identity which takes into account both its psycho-

dynamic and its psychosocial aspects. Ego identity can

be defined as









a self-structure--an internal self-
constructed, dynamic organization of drives,
abilities, beliefs, and individual history.
The better developed this structure is,
the more aware individuals appear to be
of their own uniqueness and similarity
to others and of their own strengths and
weaknesses in making their way in the world.
The less developed this structure is, the
more confused individuals seem about their
own distinctiveness from others and the
more they have to rely on external sources
to evaluate themselves. (p. 159)


Ego's Tasks in Identity Development


Marcia's definition is based on Erik Erikson's (1959,

1968) definitive work on the development of ego identity.

Erikson uses the term ego identity in recognition of the

ego's three major tasks in identity development.

The ego synthesizes childhood identifications, keeping

those that continue to fit as the adolescent moves towards

adulthood. Josselson (1980) emphasizes that "much of the

process of identity formation that takes place is selective

repudiation of possible selves. Identity is exclusive; it

is manifested in commitment and in the giving up of

potentialities: 'I will do (be) this and not that!'"

(p. 202). This melding of childhood identifications with

the developing personality and with the asirations of the

adolescent provides an inner stability for the individual,

"an assured sense of inner continuity and social sameness

which will bridge what he was as a child and what he is

about to become" (Erikson, 1959, p. 111).









The second task of the ego is to integrate the drives

of the id with the social ideologies of the super ego and

the social aspirations of the ego ideal (Erikson, 1968).

The ego's functions are to reality test the sometimes rigid

and vindictive rules of the super ego and to mitigate their

punitive nature. The ego ideal, according to Erikson (1968),

is less closely tied to morality and more closely related to

the contemporary ideals of the historical period in which

one is born than the superego is. It influences the

developing ego with the cultural values of what one should

be.

The final task of the ego in identity development

is to affirm the development of the individual within a

social reality. "The sense of ego identity, then, is the

accrued confidence that one's ability to maintain inner

sameness and continuity (one's ego in the psychological

sense) is matched by the sameness and continuity of one's

meaning for others" (Erikson, 1959, p. 89). This confirma-

tion of who one is by society is an important contributor

to a strong sense of self-esteem. It grows out of the

conviction that one is and has been performing effectively

within the expectations of that society.


Seven Perspectives on Ego Identity


In his review of the research on ego identity, Bourne

(1978a) identified seven perspectives from which the










construct, ego identity, could be explored. Erikson's

writing on ego identity encompasses all seven theoretical

perspectives. The seven perspectives are 1) structural,

2) subjective, 3) genetic (developmental), 4) dynamic,

5) adaptive, 6) psychosocial, and 7) existential.

The structural dimension emphasizes the organizing

coherent sense of self that emerges from a clear self-

definition. Marcia's (1980) definition, quoted earlier,

emphasizes this self-structure which allows an individual

to rely on an inner sense of direction and evaluation.

The next five perspectives are all contained in

Erikson's description of the ego's tasks in identity develop-

ment. The developmental perspective emphasizes the adoles-

cent's task of resynthesizing one's childhood identifications

in light of continuing developmental changes. The subjective

perspective stresses the individual's sense of continuity

that results from the blending of childhood identifications

with the new emerging aspects of the personality. The

dynamic perspective underlines the long-term developmental

processes that encompass ego identity. Different issues

affect the individual's sense of self at different points

of development. Grotevant, Thorbecke, and Meyer (1981)

underscore this aspect of developing identity. "Identity

researchers, who have almost exclusively studied college

students, are now beginning to ask which domains and which









issues within a domain are salient at different points

within a life cycle" (p. 36).

The adaptive, psychosocial, and existential perspec-

tives all accentuate the interactions between the developing

individual and his interpersonal and historical context.

Individuals must develop basic life commitments that, to

some extent, define them (existential perspective). The

self definitions are developed within a social context

(adaptive perspective), and the self-definition must be

acknowledged socially (psychosocial perspective).

According to Bourne (1978a), it is this sense of reci-

procity with society that differentiates ego identity from

the most traditional concepts of the self.


The self, in one way or another, is conceived
as the sum total of an individual's reflec-
tions, perceptions, and cognitions of him-
self, which may or may not all be derived
from social comparison, and may or may
not be entirely conscious. On the other
hand, ego identity is not simply a config-
uration of intrapsychic, self-representa-
tions, but a sense of oneself defined in
terms of a particular relationship to a
certain group, community or society.
(p. 227)


It is this perspective from which the individual affects

and is affected by his/her social surroundings that helps

to differentiate between the ego itself and ego identity.

The ego carries on reality testing functions as it integrates

the desires of the id with the rules of the super ego and

ego ideal. Ego identity can be conceived of as being more









closely tied to reality "in that it would test, select,

and integrate the self images derived from the psychosocial

crises of childhood" into a configuration which, ideally,

matches the way important others see the individual (Bourne,

1978a, p. 226).


Early Attempts to Measure Differences in Ego
Identity Development


Early attempts to operationalize ego identity utilized

Q sorts, self-report questionnaires, and semistructured

interviews. Different definitions of ego identity are

evident in the different methods used to operationalize

the concept (Bourne, 1978a; Marcia, 1980).

Bronson (1959) operationalized his conception of the

construct, identity, in terms of four "subdimensions" of

identity:

1. Degree of certainty about the subject's past and

present ideas about self.

2. The subject's certainty about his/her dominant

personal characteristics.

3. The subject's report of fluctuations in his/her

feelings about self.

4. The subject's "level of internal tension."

Semistructured interviews were rated in terms of these

subdimensions. Bronson found a modest correlation among

the four which he took as evidence of construct validity.

Bourne (1978a) reports some problems with this conclusion.









The four were arbitrarily selected and do not reflect the

scope of Erikson's concept. Also, no evidence of

discriminant validity was reported, and consruct validity

relies on both convergent and discriminant evidence.

Block (1961) and Heilbrun (1964) also based their

conceptualizations of ego identity on a sense of inner

continuity. Role variability, as perceived by the self

and observed by others, was thought to define ego identity.

Block (1961) had his subjects sort 20 adjectives with respect

to the subject's relationships with eight different indi-

viduals, such as a parent of the same sex, a close friend

of the same sex, or an acquaintance whom they would like

to know better. He hoped to find a curvilinear relationship

--i.e., individuals with excessively high and low role

stability would display maladjustment. He found that extreme

role variability did appear related to maladjustment as

measured by the Psychoneuroticism Scale of the California

Psychological Inventory (Gough, 1957). However, extreme

role stability was not indicative of neurotic qualities.

Heilbrun (1964) related role consistency to masculinity

and femininity. He found that male adolescents whose behav-

ior tended to conform to cultural stereotypes of masculinity

showed higher role consistency (using a Q sort measure)

than less masculine males. Females who were either high

or low in femininity were more consistent than females









who measured moderately feminine on The Adjective Check

List (Gough and Heilbrun, 1965).

Gruen (1960) used a Q sort (Butler and Haigh, 1954)

to measure differences between descriptions of real self

and ideal self. The hypothesis was that achievement of

a stable sense of identity would be accompanied by a reduc-

tion in the difference between idealized and realistic

self-descriptions. He found a significant positive relation-

ship between real-ideal discrepancies and willingness to

accept a false personality sketch of themselves. His con-

clusion was that both a large discrepancy between real

and ideal selves and a tendency to let others define them

were indicators of a weak sense of identity.

Hauser (1972) used a Q sort of self-descriptive state-

ments to measure "structural integration" and "temporal

stability" of identity. The statements were sorted according

to eight different types of self-images, such as "how I

am now," "how I will be in 10 years." The study was con-

ducted over a three-year period. The correlation among

the sorts within one year was defined as the "structural

integration of identity." The intercorrelation between

two sorts of the same type over two or more successive

years defined the "temporal stability" of identity. He

found differences between black and white males on both

measures.










Rasmussen (1964) designed an ego identity questionnaire

with subscales to measure the degree of resolution of each

of Erikson's first six psychosocial crises. The scale

consists of 72 items--12 items indicating healthy and

unhealthy resolutions of each of the six stages. He

presented some evidence as to convergent and discriminant

validity of his measure of ego identity. However, Bourne

(1978a) points out that Rasmussen presents very little

evidence beyond face validity to demonstrate that what

his test items tap is germane to Erikson's construct as

opposed to some general personality construct, such as

self-esteem or social competence.

Constantinople (1969) used a questionnaire based on

a Q sort by Wessman and Ricks (1966) to measure personality

development among four different grade levels of college

students. Significant differences were found between fresh-

men and seniors of both sexes on industry, inferiority

and identity and on identity diffusion for males. However,

results appear to be influenced by social desirability

problems when used with males (Tan, Kendis, Fine, and Porac,

1977).

Dignan (1965) also constructed a self-report measure,

the 50-item Ego Identity Scale, in her effort to relate

identity resolution in females to maternal identification.

The items were based on a "description of ego identity

found in the writings of Erikson and others who have









critically examined and developed the construct" (p. 478).

The components which were considered indicators of ego

identity were the following: sense of self, uniqueness,

self-acceptance, responsiveness to role expectations of

important others, stability, goal directedness, and ability

to permit involvement and intimacy with others. Dignan

presented evidence of internal consistency and stability

and some evidence for construct validity. She found that

females who scored high on the Ego Identity Scale also

reported high levels of maternal identification.


Marcia Interview


The early attempts to operationalize ego identity

investigated characteristics which would result from achiev-

ing a strong sense of identity, such as sense of inner

continuity, self-acceptance, and the ability to achieve

intimacy with others. James Marcia (1964) proposed that

a measure of ego identity should deal with the psychosocial

criteria (experiencing a "crisis" period of exploration

and developing firm commitments) that determine the degree

of ego identity achieved.

The vast majority of research on ego identity since

1966 has utilized the Marcia Semi-Structured Interview

as a measure of ego identity. Marcia's work (1964, 1966)

is closely tied to Erikson's conceptualization of ego iden-

tity development in that Marcia's statuses are based on










the presence or absence of a period of "crisis" (a time

of exploration among different alternatives), and the

formation of commitments in the process of ego identity

development. His statuses and their relationship to the

crisis and commitment variables can be outlined as below.





Achievement Moratorium Foreclosure Diffusion


Crisis
(Exploration) High High Low Low


Commitment High Low High Low




Marcia enlarged on Erikson's dichotomy of identity

achievement vs. identity diffusion by proposing two other

statuses. Youths in the moratorium status are presently

in the midst of a crisis period, but they have made no

firm commitments. They resemble diffusions in their lack

of commitments, but they are struggling for self-definition

while diffusions remain largely uninvolved with the strug-

gle. Foreclosures resemble achievements in their firm

commitments, but they have not actively explored to define

themselves. Instead, they have incorporated others' expec-

tations into their identity. Because they have not ques-

tioned the goals or assumptions their commitments are based

on, they are assumed to have acquired less self-knowledge










and flexibility regarding their occupational goals and

ideological commitments.

Marcia considered those individuals who had gone through

a period of exploration and had developed firm commitments

to self-chosen occupational and ideological goals (Identity

Achievement Status) to be the most mature. The period

of exploration was called a crisis period by both Erikson

(1959, 1968) and James Marcia (1964, 1966) because it

involves a period of doubt and indecision during which

the youth struggles with competing alternatives. The process

of struggle should, theoretically, produce ego development

as it clarifies values and goals and increases self-

knowledge. The commitments that follow from such a

decision-making period should involve strong investment

as they arise from a struggle to assess oneself and to

accurately perceive one's alternatives.

Marcia (1980) describes three advantages of the four

Identity Statuses as opposed to Erikson's (1959, 1968)

dichotomy of identity achievement vs. identity diffusion.

1. The four statuses provide for a greater variety

of approaches in dealing with identity issues.

2. There are both healthy and unhealthy aspects to

each status so the statuses describe whole people with

both positive and negative qualities. Achievements appear

to be strong, self-directed, and highly adaptive, but an

individual may settle on an identity that has limited










adaptive value for the future. "Foreclosures may be seen

either as steadfast or rigid, committed or dogmatic, cooper-

ative or conforming; Moratoriums may be viewed either as

sensitive, or anxiety ridden, highly ethical or self-

righteous, flexible or vacillating; Identity Diffusions

may be considered either carefree or careless, charming

or psychopathic, independent or schizoid" (1980, p. 161).

3. The behavior of individuals in each of the four

statuses is fairly easily differentiated, so individuals

can be classified into a status based on a 30-45 minute

interview.

The interview is rated by two or three raters for

the presence of a crisis period and the development of

firm commitments. On the basis of the ratings the subject

is assigned to one of the four statuses. Marcia (1980)

reports approximately 80% interrater reliability.

In addition to the Identity Status Interview, Marcia

(1964, 1966) also developed the Ego Identity Incomplete

Sentences Blank (EI-ISB). The EI-ISB produced a score

as an indicator of the overall level of ego identity achieve-

ment but did not assign the subject an Identity Status.

The subject's responses to sentence stems were scored by

three trained raters, who demonstrated 74% interrater

agreement.

The major problems with both of the above procedures

have been noted by several authors (Simmons, 1970, 1973;









Adams, Shea, and Fitch, 1979; Schilling, 1975; and Bourne,

1978b). The interview and the EI-ISB are time-consuming

to administer and score, and they rely on subjective judgment

in the ratings of the response.


Other Measures of Ego Identity


Marcia (1964, 1966) used his Ego Identity Incomplete

Sentences Blank (EI-ISB) to validate his Identity Status

Interview. He found that achievements scored the highest

and diffusions scored the lowest on this overall measure

of ego identity development. Simmons (1970) created a

24-item easily scored multiple choice measure, The Identity

Achievement Scale, based on Marcia's Ego Identity Incomplete

Sentences Blank. Because it is objectively scored, it

eliminates the need for trained raters and the possible

problems of subjective judgment used in the ratings. The

measure does not assign subjects to an identity status

but provides an overall rating of identity achievement.

Tan, Kendis, Fine, and Porac (1977) developed a 12-item

ego identity scale with a forced choice format which measures

identity achievement vs. identity diffusion. Evidence

was presented for the test's construct validity. There

is also some evidence that the scale measures achieved

identity following a period of exploration rather than

a foreclosed identity. The scale does not attempt to assign

subjects to ego identity statuses.









Adams, Shea, and Fitch (1979) constructed a 24-item

self-report measure that assigns individuals to an overall

identity status based on three content areas--Career Orien-

tation and Political and Religious Ideology. Convergent-

divergent, concurrent and predictive validity, and

test-retest reliability have all been documented in the

reported data.

Schilling (1975) devised a forced choice format for

assigning individuals to two Ego Identity Statuses. His

measure, the Ego Identity Questionnaire, consists of two

parts. Part A assesses identity based on occupational,

political, and religious concerns, while Part B assesses

identity based on interpersonal issues. Schilling provided

some evidence of the instrument's validity with males.

The limitations of both of these measures have been discussed

in Chapter 1.

Baker (1971) based his definition of identity on four

qualities he derived from Erikson's writings: "a) knows

who he is, b) knows where he is going, c) perceives himself

as having 'inner sameness and continuity,' and d) is certain

about the way his perception of himself compares to the

perceptions which others have of him" (p. 167). He con-

structed a 32-item measure consisting of 8 Likert-type

scale items for each of the four qualities. His internal

reliability coefficients (Kuder-Richardson-Formula 20)

were "quite low" (p. 170), but he achieved good









intercorrelations among three of the four areas. He con-

cluded that his measure is a successful translation of

Erikson's work on ego identity into operational terms.


Gender Difference in Ego Identity Development


Much of the research in ego identity formation has

followed Erik Erikson's (1968) conception of gender differ-

ences in identity development. Erickson emphasized commit-

ments toward occupational goals and toward a political

and religious ideology as important processes in male

identity development.

Female identity has a biological base in what Erikson

referred to as a "vital, productive inner space." Inner

space refers to the physical reality of the uterus and

vagina and also to the somatic awareness of a woman's repro-

ductive capacity. This awareness affects her manner of

relating to a male, the potential father of her children,

and her sense of direction in life. While she may experiment

with and enjoy development in "male" areas of career and

ideology, "womanhood arrives when attractiveness and experi-

ence have succeeded in selecting what is to be admitted

to the welcome of the 'inner space' for keeps" (p. 283).

While Erikson emphasized the importance of a woman's

biological inheritance on her identity development, Douvan

and Adelson (1966) highlighted the importance of culturally










influenced sex role expectations on developing males and

females.

The key terms in adolescent development
for the boy in our culture are the erotic,
autonomy (assertiveness, independence,
achievement), and identity. For the girl
the comparable terms are the erotic, the
interpersonal, and identity . what
the girl achieves through intimate connection
with others, the boy must manage by discon-
necting, by separating himself and asserting
his right to be distinct. (pp. 347-348)


Research by Josselson, Greenberger, and McConochie

(1977a, 1977b) supports the above view. They studied psycho-

social maturity in adolescent males and females and found

that high maturity females used interpersonal relation-

ships for identity resolution and were less focused than

the high maturity males on career goals as a source of

self-esteem. "They are, in a word, identity seekers,

attempting to discover who they are and who they want to

be in relation to the significant others in their lives"

(1977b, p. 159). High maturity males were more concerned

with who they would become than in how they would relate

to others.

The early studies utilizing the Marcia Interview were

conducted with males, and the interview focused on the

three areas considered relevant to male identity development:

occupational goals and political and religious ideology.

The question that underlay this research was "What are

the characteristics of late adolescent males who pursue









different modes of identity resolution?" (Marcia, 1980,

p. 162).

When research in ego identity using the Marcia Inteview

was extended to females, it was assumed that the occupational

and ideological issues considered focal for males were

not as pertinent to female identity development. Marcia

and Friedman (1970) in the first study using the Marcia

Interview with females added a section of questions assessing

"attitudes towards premarital intercourse" to the interview

format.

Both Schenkel and Marcia (1972) and Hopkins (1980)

investigated the relative importance of sexual values to

female identity development, but with conflicting results.

Schenkel and Marcia (1972) found that identity status

assigned by the sexual ideology section of the interview

("attitudes toward premarital sex") accounted for more

variance on their several dependent measures than identity

status assigned by the occupation and the political and

religious ideology sections on the interview. The religious

ideology section of the interview predicted performance

on the dependent measures almost as well as the sexual

ideology section.

Hopkins (1980) conducted two interviews with her female

subjects--the standard Marcia Interview used with males--

occupational goals and political and religious ideology

("outer space" interview) and an interview on premarital









sex, life plans, and sex roles ("inner space" interview).

Her conclusion was

The women were just as likely to have
experienced an outer space crisis as an
inner space crisis. Factor analysis revealed
that the inner space and outer space statuses
were distinct, independent factors. This
finding supports the use of separate inner
and outer space interviews in assessment
of identity status. . Regression analysis
showed that the outer space identity status
was the better predictor of the dependent
variables. The best measure of female
identity status was found to be either
the outer space interview alone or joint
use of both interviews. (p. 9)


Schilling (1975) designed a questionnaire to determine

the importance of sex role and interpersonal issues for

male identity formation. His results appear to indicate

that sexual and interpersonal issues are important to males

as well as females in clarifying their identities.

Several other investigators have studied the importance

of the different content areas to identity development

in studies using both males and females. Results have

confirmed the importance of occupational choice, sex values

and sex role, and political and religious ideology to

identity issues for both young men and women.

Matteson (1977) reports a study by Nevid, Nevid,

O'Neill, and Waterman (1974) that was the first to compare

the importance of the ORP (occupational and religious and

political ideology) and the sexual ideology interviews

("attitudes toward sexual intercourse") for subjects of










both sexes. This study investigated only the distribution

of subjects in the four Identity Statuses, assigning subjects

to a status in each of the content areas of the interview.

No dependent measures were used. Using the sexual ideology

section of the interview, a higher percentage of women

than men was assigned to statuses indicating exploration

(Achievement and Moratorium). The authors report that

there was no evidence that the other content areas (ORP)

were less important to female identity than to male

identity.

Mattesop (1977) reproduced the above study with Danish

adolescents. He used "male and female sex roles" to explore

sexual ideology rather than "attitudes toward premarital

intercourse." His results paralleled Nevid et al. (1974).

When subjects were assigned on the basis of the sex role

portion of the interview, a higher percentage of females

than males were assigned to categories involving exploration

(Achievement and Moratorium--51% of the females compared

to 38% of the males). He concluded that more females than

males appear to undergo an exploration of sexual identity.

Matteson (1977) also used continuous scales of explora-

tion and commitment to test the hypothesis of a higher

degree of exploration of sexual identity among women than

men. Exploration and commitment were separately rated

for each content area on a four-point scale: 1) limited,

2) some, 3) considerable, 4) very intense. This method







31

of measurement also indicated a higher level of exploration

of sexual identity for women than men. Matteson (1977)

reported the following results when comparing exploration

and commitment between sexes and among the four content

areas (occupational goals, political ideology, values,

and sex roles):

1. There is generally a higher level of commitment

among Danish females than males, across all content areas

except politics, in which males show a slightly higher level

of commitment.

2. There is some indication of a higher level of

exploration for women generally using either categories

or scales.

3. Females as well as males indicated high levels

of occupational exploration.

4. For both men and women the highest levels of com-

mitment occur in the areas of sex roles and values.

Matteson concluded that it may be valuable in future

research to consider the process variables (exploration

and commitment) separately and to separately analyze the

impact of each of the content areas on identity development

for males and females.

Hodgson and Fischer (1979) studied sex differences

in identity and intimacy development among college under-

graduates. The dependent variables were identity status,

intimacy status, and self-esteem. In addition to the







32
standard Marcia Interview, they included a section on overall

sexual ideology (sex values and sex roles). The authors

did not assign an overall identity status but assigned

subjects to either a high (Achievement or Moratorium) status

or a low (Foreclosure or Diffusion) status based on "part

conflicts" which were explored in the content areas of

the interview. The content areas are as follows: Occupation,

Overall Ideology, Religion, Politics, Overall Sexual

Ideology, Sexual Values, and Sex Roles. The first of

Hodgson's and Fischer's hypotheses was that there would

be more males than females in the more advanced stages

(Achievement, Moratorium) of identity development in the

areas of occupation and political and religious ideology;

and females would more commonly be in the advanced stages

of identity development in the areas of sexual ideology

and sex roles. These predictions were supported. Signifi-

cantly more males than females were developmentally advanced

in areas of occupation, politics, and religion and overall

(religious and political) ideology, while significantly

more females than males were developmentally advanced in

sex roles and overall (sex values and sex role) sexual

ideology. Sex Values area produced no significant sex

difference. Hodgson and Fischer concluded that these results

provided evidence for the influence of sex roles on identity

development. Young men, they concluded, seemed to define








33

themselves with respect to issues of competence while young

women's identity issues were focused on relating to others.

Waterman and Nevid (1977) investigated the relative

importance of premarital sexual values and of the occupa-

tional and political and religious ideology sections of

the interview for males and females. The authors predicted

that females would be more likely than males to be found

in statuses involving a crisis period of exploration

(Moratorium and Achievement) as assigned by the premarital

sexual values section of the interview. Females would

also be more likely to have undergone a crisis in the sexual

area than in the ORP area. Results indicated that females

were more likely than males to have experienced a sexual

identity crisis and males were more likely to be foreclosed

in this area. Males as well as females were more likely

to be committed about sex than any other area of the inter-

view. In all areas other than sex, female identity appeared

to have a similar pattern to male identity.

Kacerguis and Adams (1980) investigated whether for

both males and females advanced overall ego identity status

would be associated with more advanced intimacy development.

Using only three content areas (occupation and political

and religious ideology), they predicted that the occupational

and political ideology sections of the interview would

be more indicative of intimacy development for males, while

religious identity, as an affiliative, nurturant, and









34

expressive life issue, would be more predictive of intimacy

development for females. Results indicated that for both

males and females, occupational identity was the primary

predictive factor in the level of intimacy development.

Interpersonal relationships (friendship and dating

relationship) have been proposed as important areas of

assessment in ego identity development (Grotevant, Thorbecke,

and Meyer, 1982). The ability to establish intimate rela-

tionships has been described by Erikson (1959, 1968) as

the central crisis of the developmental stage following

the Identity Achievement vs. Identity Diffusion crisis

of adolescence. Grotevant et al. (1982) stress that it

is important to distinguish between the achievement of

one's identity and ideology with respect to interpersonal

relationships and the actual establishment of intimate

relationships. They suggest that exploring interpersonal

issues and making commitments to an ideology of interpersonal

relationships is an important aspect of adolescent ego

identity development and is a precursor to successfully

meeting the challenge of the next stage of development.

Grotevant et al. (1982) extended the Marcia Interview to

include friendship, dating, and sex roles as important

areas of assessment and provided evidence that the extended

interview is psychometrically sound.

Thorbecke and Grotevant (1982) used the extended inter-

view to assess gender difference in interpersonal identity










formation. Measures of Vocational Identity, Masculinity-

Femininity, and Achievement Motivation were the dependent

variables. The data indicated that young women who were

more advanced in vocational identity were also more advanced

in interpersonal identity formation. Their vocational

identity scores were correlated with friendship commitment,

dating exploration, and dating commitment. For males,

vocational and interpersonal identity achievement appeared

to proceed more independently. Young women were signifi-

cantly more identity achieved than men in the friendship

domain although there were no gender differences in the

dating domain. Finally, for both males and females, expres-

sive attributes of psychological femininity were positively

related to interpersonal identity formation.

These data support the conclusion that for females

the processes of interpersonal and vocational identity

formation may be more interrelated than for males. Thorbecke

and Grotevant (1982) suggest that the late adolescent female

may be in a "double bind dilemma"--she is attempting to

negotiate her occupational and interpersonal identities

simultaneously.


While most young men are clearly socialized
for career establishment, women are in
the more complex position of being socialized
for both career and family establishment.
The perceived need to integrate these two
roles may make identity formation in these
two domains more interrelated for females
than for males. (p. 489)









Separating Content Areas


Much of the research on ego identity development has

assumed that identity status for one individual is similar

across different content areas (Bourne, 1978b). This assump-

tion allows researchers to assign an individual to one

overall identity status regardless of the number of areas

that are assessed. Grotevant et al. (1982) suggest that

as the number of areas considered in assigning an Identity

Status increase (i.e., the addition of friendship, dating,

and sexual values areas), an overall rating becomes less

meaningful. Matteson (1977) and Grotevant et al. (1982)

argue that adolescents undergo a series of crises and that

one area of their lives may be stable while another area

is very much in crisis. Overall ratings may mask the

processes of identity development rather than clarify them.

Coleman's (1978) work with 800 boys and girls ages

11 to 17 supports the idea of a different sequencing of

consideration of identity domains among adolescents. His

"focal theory . proposes that at different ages particu-

lar relationship patterns come into focus, in the sense

of being most prominent, but that no pattern is specific

to one age only" (p. 8). Issues are not necessarily linked

to a particular age, nor is their sequence fixed. This

process of focusing on different issues at different times,

he suggests, spreads the process of adaptation over a span









of years and contributes stability to the adolescent time

of stress and change.

Schenkel (1975) investigated the content areas of

the Marcia Identity Status Interview for females by

separately assigning identity statuses for occupational,

political ideology, religious ideology, and sexual values

areas of identity. She found wide discrepancies in some

of her subjects across content areas.

Waterman and Waterman (1971) and Waterman, Geary,

and Waterman (1974) separately assigned college males to

identity statuses in terms of two content areas--ideology

(religious and political) and occupation. Their data indi-

cated that college males first confronted decisions regarding

occupations and only later explored areas of ideology.

The sexual values content areas was not included on these

interviews.

Marcia (1980) described ego identity as a self-structure

that gets formed little by little. Each decision builds

on previous experience and contributes to future decisions

until a consistent whole evolves. Hypothesizing that "the

area in which an individual is most involved may vary from

person to person or over time within an individual" (p. 328),

Rogow, Marcia, and Slugoski (1983) investigated the relation-

ship between identity status within a content area and

an overall identity status with male subjects only. Expand-

ing the standard Identity Status Interview (Occupational









and Political and Religious ideology) to include two new

interpersonal-sexual areas ("attitudes toward sexual expres-

sion" and "sex role beliefs"), they described three aims

of the study:

1. To assess the concurrence of identity statuses

across content areas.

2. To assess the relative importance of any one content

area to overall identity status.

3. To determine the relationship between identity

status in a content area considered to be important by

the subject and the subject's overall identity status.

The results suggested that the process variables of

crisis and commitment may be more important than any one

particular content area in investigating ego identity devel-

opment. The ideology areas of the interview appeared to

correspond to the overall status rating more than the occupa-

tional area. The ideological areas were also more predictive

of results on the dependent variables than were the other

areas. However, all interview areas corresponded well

with the overall identity status and discriminated signifi-

cantly on the dependent variables. Result on the "attitudes

toward sexual expression" and "sex role beliefs" areas

indicated that interpersonal sexual concerns are important

for male identity development as well as for female identity

development.









Exploration and Commitment Variables


Matteson (1977) and Grotevant et al. (1982) have both

underlined the importance of the variables of crisis and

commitment in their work on ego identity development.

In addition to assigning subjects to identity statuses,

they both measured the processes of exploration and the

development of commitments in the different areas of iden-

tity. "The advantage of rating exploration and commitment

separately is that the researcher then has two continuous

scales available for each identity domain; these can be

used in correlational analysis with other variables"

(Grotevant et al., 1982, p. 45). Looking at the variables

of exploration and commitment separately allows researchers

to explore many unresolved issues in identity formation.

How are these two processes related? How do adolescents

explore the different areas? Is there a sequence for ado-

lescents as they explore the different identity content

areas? Does commitment in one area provide stability for

exploration in other areas? Are there sex differences

in the patterning of exploration and commitment across

content areas?

Matteson (1977) concludes


If one were designing research on the effects
of three factors in the identity process
(search among alternatives, degree of
involvement, and timing of commitments)
upon the outcomes of that process, it would
be preferable to measure each of these









factors separately, and to measure the
variables in degrees along a quantitative
scale, rather than lump persons into two
categories on each variable. The development
of science proceeds from categorization
to the isolation of specific variables
and the discernment of their interactions.
(p. 355)


The main purpose of this study is to create an objective

measure of ego identity, the Exploration and Commitment

Scale, that assigns individuals to an identity status in

each content area of identity based on test items that

clearly reflect the presence or absence of exploration

and commitment. Separate analyses of the variables of

exploration and commitment will be enabled by constructing

test items for each area that reflect either exploration

or commitment.

For the present study I have also constructed a Self-

Report Measure of Exploration and Commitment in which the

subject is asked to assess and report separately his/her

level of exploration and level of commitment in each of

the following areas: career, dating, friendship, political

ideology, religious ideology, sexual values, and sex role

values. The subjects' self-reports can then be compared

to their responses to the exploration items and to the

commitment items on the Exploration and Commitment Scale.









Social Desirability


One of the criticisms of assessments which rely on

interviewing is that the rating of the interview responses

is affected by the subjectivity of the raters. A problem

of equal magnitude in test construction is the problem

of social desirability. "People tend to give responses

that are socially desirable, responses that indicate or

imply approval of actions or things that are generally

considered to be good" (Kerlinger, 1973, p. 486). Test

items need to be constructed so that the socially desirable

direction of response is minimized. Anonymity in responding

helps decrease the effects of the social desirability

motive. However, an assessment of the possible social

desirability effects on responses in any new measure is

an essential part of providing evidence of reliability

and validity.


Instrument Development--Reliability and Validity


Since a major portion of this study is the creation

of an objective measure of ego identity, it is important

to review methods of assuring an instrument's reliability

and validity. Reliability concerns the extent with which

measurements are repeatable under the same conditions,

while the validity of measurements has to do with their

generalizability (Nunnally, 1970). Three measures of

reliability described by A.P.A. (1954) are the 1) coefficient









of internal consistency, 2) the coefficient of equivalence,

and 3) the coefficient of stability.

The coefficient of internal consistency is a reliability

estimation based on homogeneity, or the amount of correlation

among items on a single test. Coefficient alpha is generally

considered to be the most useful equation for testing

reliability based on internal consistency. The Kuder-

Richardson formula 20, a short-cut version of that equation,

is also commonly used (Nunnally, 1970).

Using alternate forms of the same test or correlating

part scores for different parts within the same test also

provide evidence of reliability--usually referred to as

the coefficient of equivalence (Nunnally, 1978).

The coefficient of stability refers to the correlation

between test and retest scores with an intervening time

period. A period of several weeks is suggested to diminish

the effects of memory on the second set of scores obtained

(Nunnally, 1970).

The four types of validity most commonly referred

to in the literature are described by the A.P.A. (1954).

These are 1) content validity, 2) construct validity, 3)

concurrent validity, and 4) predictive validity.

Content validity depends upon the adequancy with which

the test items represent the entire content domain about

which conclusions are to be drawn (A.P.A., 1954). Nunally

(1970, 1978) describes two standards for ensuring content









validity: a) a representative collection of test items,

and b) sensible methods of test construction. Evidence

that the above standards have been met is provided by a

thorough and logical explanation of the plan and procedures

of this construction. The values that underlie the test

should be made explicit, and it should be indicated how

these values guided in formulation of the test outline

and the selection of items. "Even though there are often

problems involved in ensuring content validity, inevitably

content validity rests mainly on appeals to reason regarding

the adequacy with which important content have been sampled

and on the adequacy with which the content has been cast

into the form of test items" (Nunnally, 1970, p. 137).

Two statistical measures provide some circumstantial evidence

for content validity. The test should demonstrate at least

a moderate level of internal consistency (see reliability

measures above). Correlating scores on different instruments

which purport to measure the same thing also provide some

indirect evidence for content validity.

Both predictive and concurrent validity concern cor-

relating scores on an instrument with scores obtained for

some criterion measure. Nunnally (1978) combines predictive

and concurrent validity since they are both criterion related

measures only differing in the timing of the data collection

for the criterion measure. Predictive validity refers

to correlation with a subsequent criterion measure while









concurrent validity refers to correlation with scores

obtained at the same time. The nature of the problem dic-

tates when the two sets of measures will be obtained.

Criterion related validity is another term often used for

this type of validity measurement.

Construct validity is a measure of the meaning of

test scores in terms of psychological "constructs" (Cronbach

and Meehl, 1955). Nunnally (1978) describes three major

aspects of the construct validation process.

1. The domain of observables related to the construct

should be specified.

2. Using empirical research or statistical analysis

determine the extent to which the observables measure the

same things or different things.

3. Studies (or controlled experiments) should be

performed to determine if results that are predictable

from well-grounded hypotheses concerning the construct

can be obtained.

In specifying the domain of observables Nunnally (1978)

suggests explicitly outlining the theory of how the construct

can be translated into measurable variables and how these

variables relate to one another. Outlining the domain

in this way is helpful in proceeding with the next step

which involves determining how well the measure of observ-

ables "go together" in empirical studies. Correlation

between scores on measures which are related to the construct










are computed. If the measures correlate highly with one

another, it can be concluded that they measure the same

thing. Measures of internal consistency also provides

some evidence for construct validity.


To the extent that the elements of a domain
show this consistency, it can be said that
some construct may be employed to account
for the data, but it is by no means neces-
sarily legitimate to employ the construct
name which motivated the research. In
other words, consistency is a necessary
but not sufficient condition for construct
validity. . Sufficient evidence for
construct validity is the fact that the
supposed measures of the construct (either
a single measure of observables or a com-
bination of such measures) behave as
expected. (Nunnally, 1978, p.146)


Construct validity is supported by evidence of predic-

tive validity and content validity. An instrument which

is designed to measure a construct can accurately serve

as a predictor of other scores or behavior which would

enhance the contract validity of the measure. The same

procedures intended to ensure content validity are helpful

in defining the domain of observables in construct validity,

since content validity depends on the thoroughness with

which the domain of observables has been sampled and

converted into good test form.

Nunnally (1978) concludes discussion of this complex

area as follows:


One could rightly argue that all this fuss
and bother about construct validity really









boils down to something rather homespun--
namely, circumstantial evidence for the
usefulness of a new measurement method.
New measurement methods, like most new
ways of doing things, should not be trusted
until they have proven themselves in many
applications. (p. 109)


Hypotheses


The current study provides some evidence to support

the usefulness of an instrument, the Exploration and

Commitment Scale. This scale assesses the amount of explora-

tion and commitment in four content areas considered relevant

to identity development: Career Values, Political and

Religious Ideology, Sexual and Sex Role Values, and the

Interpersonal Areas of Dating and Friendship.

The Exploration and Commitment Scale was designed

to provide the following information.

1. An overall score which indicates the level of

identity achievement but which does not categorize an

individual according to method of identity resolution (i.e.,

foreclosed or achieved) can be calculated.

2. Exploration test items indicate the amount of

exploration the individual has undergone or is currently

experiencing. An exploration score can be given for each

area, or an overall score can be computed.

3. Commitment test items indicate the amount of com-

mitment the individual currently is expressing. A commitment









score can be given for each area, or an overall score for

commitment can be computed.

4. Individuals can be categorized into one of Marcia's

four Identity Statuses in each content area. An overall

Identity Status could be calculated but is not recommended.

As the number of content areas used to define ego

identity increases, the idea of an overall Identity Status

becomes less meaningful. The interaction of exploration

and commitment within each content area and the pattern

that develops within the individual as he or she consolidates

an identity in the content areas define the processes of

ego identity development. Assigning an overall status

may mask the processes rather than clarify them.

The reliability of the Exploration and Commitment

Scale was assessed by measures of internal consistency

(Coefficient Alpha, Nunnally, 1970), and test-retest stabil-

ity (Pearson Correlation, Agresti & Agresti, 1979). Evidence

is presented regarding content, construct and, concurrent

validity.

In assessing content validity, one must consider whether

the test items adequately sample the entire domain of the

construct and whether the domain is accurately represented

in the test items. In Chapter 3, the methods used in test

item selection and the relationship of the individual test

items and content areas to Erikson's construct of ego iden-

tity are discussed. Erikson's writing (1959, 1968) and









the body of research literature exploring ego identity

development support the use of the processes of exploration

and commitment in operationalizing ego identity. The four

content areas of the Exploration and Commitment Scale are

the areas emphasized in the research literature, within

which the two processes are explored.

Whether the test items do indeed reflect these two

processes is also important to establish. The Pearson

Correlation (Agresti & Agresti, 1979) was used to test

the relationship between exploration scores on the Explora-

tion and Commitment Scale and exploration scores on the

Self-Report Measure of Exploration and Commitment (Appendix

A). The Pearson Correlation (Agresti & Agresti, 1979)

was also used to test the relationship between commitment

scores on each instrument. The Self-Report Measure of

Exploration and Commitment (Appendix A) provides an indica-

tion of the individual's assessment of his (her) own level

of exploration and commitment.

Correlating the scores on these two instruments provides

evidence that the test items do reflect the individual's

experience of exploration and commitment. Two hypotheses

test the relationship between the Exploration and Commitment

Scale and the Self-Report Measure of Exploration and Commit-

ment (Appendix A).









Hypothesis 1


There will be a significant positive relationship

between overall exploration scores on the Exploration and

Commitment Scale and total exploration scores on the Self-

Report Measure of Exploration and Commitment (Appendix A).


Hypothesis 2


There will be a significant positive relationship

between overall commitment scores on the Exploration and

Commitment Scale and total commitment scores on the Self-

Report Measure of Exploration and Commitment (Appendix

A).

Correlating test scores on other instruments that

purport to measure the same construct is another indirect

piece of evidence for content validity. The Pearson correla-

tion (Agresti & Agresti, 1979) was used to assess the rela-

tionship between the overall score on the Exploration and

Commitment Scale and the total score on the Simmons (1973)

Identity Achievement Scale. Both instruments produce a

total score which reflects the level of identity achievement

attained. Additionally, a comparison of the mean scores

on the Simmons' test of two groups--l) those scoring above

the median on the Exploration and Commitment Scale and

2) those scoring below the median on the Exploration and

Commitment Scale--was accomplished using the Kruskal-Wallis








50

test. This comparison provides evidence that both measures

significantly differentiate among a range of scores. Two

hypotheses test the relationship between the Exploration

and Commitment Scale and Simmons (1970) Identity Achievement

Scale.


Hypothesis 3


There will be a significant positive relationship

between the overall scores on the Exploration and Commitment

Scale and total scores on the Simmons (1970) Identity

Achievement Scale.


Hypothesis 4


There will be a significant difference in mean scores

on the Simmons (1970) Identity Achievement Scale between

those who scored above the median and those who scored

below the median on the Exploration and Commitment Scales.

Construct validity is the most difficult to describe

and to assess. Specifying the domain of items used to

operationalize the construct is one step in providing evi-

dence that a measure's results have a valid meaning in

terms of the psychological constructs it is said to rep-

resent. So, evidence that an instrument has good content

validity supports the measure's construct validity.

Construct validity also addresses the question of

whether the test items combine to measure a coherent









psychological construct. Evidence of internal consistency

(Coefficient Alpha, Nunnally, 1970) provides some support

for the coherence of items within a test. Convergence-

divergence data can also support the construct validity

of a measure. The internal consistency of the overall

scale as well as of the exploration items and the commitment

items was assessed using the Coefficient Alpha (Nunally,

1970). Evidence that the exploration items correlate highly

and the commitment items correlate highly and that both

combine into a scale with an adequate level of internal

consistency supports the construct validity of the instru-

ment. Additionally, the convergence of the exploration

and commitment items within each of the four content areas

and their divergence from the other scale totals was assessed

using multiple Pearson correlations (Agresti & Agresti,

1979). High correlations of test items within a content

area and low correlations with other content area totals

provides evidence that the four content areas measure dif-

ferent aspects of ego identity.

Construct validity is supported by evidence of concur-

rent or predictive validity. The central question in con-

current predictive validity is whether subjects score on

a measure as one would predict they should. Predictions

can be theoretically based or based on the pattern of previous

research results. Studies which have examined ego identity

development within content areas have typically compared









male and female differences in exploration and commitment

within the content areas. Six studies (Matteson, 1977;

Waterman & Nevid, 1976; Hodgson & Fischer, 1979; Nevid

et al., 1974; Kacerguis & Adams, 1980; Grotevant et al.,

1982; and Thorbecke & Grotevant, 1982) have investigated

male-female differences within the content areas. The

pattern of results from these studies appears to indicate

that females generally have a higher level of exploration

and commitment in the Interpersonal and Sexual content

areas. There are some data to support the hypothesis that

the Ideology content area (particularly political ideology)

is more important to males' identity development than to

females' (Matteson, 1977; Nevid et al., 1974; Waterman

& Nevid, 1976; and Hodgson & Fischer, 1979). Results in

the Career Content area from the six studies have been

mixed indicating a high level of salience to both males

and females. Accordingly, evidence of concurrent validity

is presented by testing the following hypotheses:


Hypothesis 5


The mean exploration score for females for the Sexual

Content area of the Exploration and Commitment Scale will

be higher than the mean exploration score for males in

the sexual content area.









Hypothesis 6


The mean commitment score for females for the Sexual

Content area of the Exploration and Commitment Scale will

be higher than the mean commitment score for males in the

Sexual Content area.


Hypothesis 7


Females will be significantly more likely than males

to be categorized in the more advanced ego identity statuses

(Achievement and Moratorium) in the Sexual Content area.


Hypothesis 8


The mean exploration score for females for the Inter-

personal Content area of the Exploration and Commitment

Scale will be higher than the mean exploration score for

males in the Interpersonal Content area.


Hypothesis 9


The mean commitment score for females for the Inter-

personal Content area of the Exploration and Commitment

Scale will be higher than the mean commitment score for

males in the Interpersonal content area.


Hypothesis 10


Females will be significantly more likely than males

to be categorized in the more advanced ego identity statuses









(Achievement and Moratorium) in the Interpersonal Content

area.


Hypothesis 11


The mean exploration score for males for the Ideology

Content area of the Exploration and Commitment Scale will

be higher than the mean exploration score for females.

One of the criticisms of objective measures is that

the responses are influenced by the subject's desire to

appear in a socially acceptable light. To provide evidence

that the responses on the Exploration and Commitment Scale

are not confounded by the social desirability motive, the

overall scores were correlated with the overall scores

on the Crowne and Marlowe (1964) Social Desirability Scale.

The following hypothesis was tested.


Hypothesis 12


There will be no significant relationship between

the overall scores on the Exploration and Commitment Scale

and total scores on the Crowne and Marlowe (1964) Social

Desirability Scale.















CHAPTER III
METHOD


Subjects


All subjects were undergraduate students at the

University of Florida. A total of over 400 students

participated in various stages of this research. Over

100 students in Elementary Education and Introductory

Psychology classes participated in pilot studies with

earlier forms of the Exploration and Commitment Scale.

Their efforts were valuable in the creation of the final

instrument. Two hundred and seventy-one students in

Introductory Sociology and Introductory Psychology courses

completed the series of tests necessary to test the reliabil-

ity and validity of the final version of the Exploration

and Commitment Scale. Data from 26 of these subjects

were not used because the subjects were overage or because

they had incompletely filled out the forms. The final

sample of 245 consisted of 95 males and 150 females between

the ages of 18 and 23 years. The mean overall age was

19.4 years. For females the mean age was 19.5 years;

for males, it was 19.2 years. This age range was chosen

as this is the developmental period when males and females

are typically addressing the identity issue.

55









Procedure


Development of the Exploration and Commitment Scale


The basic assumption in operationalizing the construct

of ego identity was that the variables of exploration

and commitment were essential in its definition. Both

Erik Erikson (1959, 1968) and James Marcia (1964, 1966)

emphasized these two variables in describing the processes

of ego identity development. Others (Matteson, 1977;

Waterman & Nevid, 1977; and Grotevant, Thorbecke, & Meyer,

1981) have also selected these as essential factors in

operationalizing the ego identity construct. The vast

majority of the research on ego identity development in

the last 20 years has utilized the Marcia Interview.

The Marcia Interview is aimed at assessing the amount

of exploration the subject has engaged in and his or her

degree of expressed commitment toward personal goals and

values. Exploration and commitment on the Exploration

and Commitment Scale were defined as Marcia (1964) defined

them. Exploration is "a time during adolescence when

the individual seems to be actively involved in choosing

among meaningful alternatives. . Commitment refers

to the degree of personal investment the individual expresses

in a course of action or belief" (p. 24).

Thirty-two statements were constructed which indicate

past or current exploration. The statements express









involvement in the process of examining one's beliefs,

opinions, or positions. Past exploration is indicated

by a change in opinions, beliefs, or positions over time

or by an increased ability to consider a broader range

of options.

Thirty-two commitment items were written reflecting

personal investment in a belief, or in a position and/or

a feeling of stability or consistency in the position

taken. No effort was made to differentiate between fore-

closed commitment and commitment made after an exploration

of alternatives. The presence of an expressed sense of

a firm stance with regard to an issue, however arrived

at, underlies the commitment items.

Each of the exploration and commitment variables

was considered within the context of four areas of an

individual's life which have been considered relevant

in research on ego identity development in late adolescence:

1. career goals,

2. ideology (political and religious),

3. sexual and sex role values,

4. interpersonal (friendship and dating values).

The resulting format consisted of four subtests,

each having eight exploration and eight commitment items.

Content areas and examples of exploration and commitment

items for each area follow. The Exploration and Commitment

Scale is entirely reproduced in Appendix B.









Career Exploration Items


1. Trying to narrow my interests down to one career

has been a struggle for me.

57. Talking over career possibilities with others has

been helpful to me.


Career Commitment Items


6. I consider myself settled as far as my career plans

are concerned.

10. The career area that I have chosen suits my interests

and abilities.


Ideology Exploration Items


5. I have gone through (or am now going through) a period

of questioning and uncertainty about my religious

beliefs.

23. Having political discussions with friends has helped

me to clarify what I think politically.


Ideology Commitment Items


2. I have strong political opinions.

12. I feel like my religious beliefs are an important

influence in my life.









Sexual Exploration Items


27. Dating different people has influenced me to rethink

some of my ideas on men's and women's roles in marriage.

35. I've had to figure out for myself what is right for me

sexually.


Sexual Commitment Items


42. While my friends may not agree with my sex role values

(traditional or liberal), it would not be easy for

them to change my mind.

50. I feel strongly about what I think is right and wrong

in the sexual area.


Interpersonal Exploration Items


11. What I would look for in a person I would date has

changed over the years since I first started dating.

15. I've worked at understanding my friends better.


Interpersonal Commitment Items


4. I can tell after the first few dates if someone is

right for me.

8. I don't easily change my ideas about friends.


Data Collection


Pilot studies on three earlier versions of the Explora-

tion and Commitment Scale were conducted with over 100









undergraduates in Elementary Education and Introductory

Psychology classes. Items were refined following each

study. The resulting test items were examined by two

advanced graduate students and a clinical psychologist

who had been trained in assessing exploration and commitment

items according to Marcia's definition. The final test

version incorporated their suggestions.

Two hundred and seventy-one Introductory Sociology

and Introductory Psychology students were asked to complete

the final version of the Exploration and Commitment Scale

(Appendix B) along with the following instruments:

1. The Self-Report Measure of Exploration and

Commitment (Moberly, 1985) (Appendix A).

2. The Social Desirability Scale (Crowne and Marlow,

1964) (Appendix C).

3. The Identity Achievement Scale (Simmons, 1970)

(Appendix D).

A subsample of these students (N=45) completed a

retest of the Exploration and Commitment Scale after a

one-week interval.

The data were collected by the author during the

regular class period. Since all the classes had at least

75 students, the forms were filled out in large groups.

The data were collected anonymously except for the sample

of students who were in the test-retest group. Their

social security numbers were removed from their tests









as soon as the test and retest data were matched. The

students were informed of the nature of the study by the

following paragraph on the front of their packet of tests.


General Instructions for Participants in this Study


This is a study exploring how each
of you decides your position on issues,
and how each of you thinks and feels. A
series of several different kinds of
tasks will be used to help gain an
understanding of your values, thoughts
and feelings. There are no right or
wrong answers, only your answers. All
the data will be handled on a coded
basis and reported as group data. We
are asking for your age, sex, and social
security number to classify and code the
data. Following the collection of all
the data, your social security number
will be deleted. Thank you for par-
ticipating in this study.



The students were reminded to answer all questions

and to fill in their age and sex on the front of the packet.

Those not in the test-retest group were told not to fill

in the social security number. The students in the test-

retest group were given 10 extra points toward their overall

grade in the Introductory Psychology course for their

participation. No other compensation was given to any

of the participants.









Instruments


The Exploration and Commitment Scale was constructed

for the purposes of this study. The Scale consists of

64 statements, and the subject is asked to indicate on

a 7-point scale how much the statement agrees with the

way he or she thinks and/or feels. The analysis and valida-

tion of this instrument is the primary goal of this study.

The data relating to its reliability and validity will

be considered in Chapter 4.

The Exploration and Commitment Scale was constructed

to provide several different kinds of information.

1. A total score (the sum of scores from all test

items) which indicates the overall level of both exploration

and commitment. This score can be considered an indicator

of ego identity achievement.

2. A total exploration score (the sum of all odd-

numbered items) which indicates the amount of exploration

the individual has undergone or is currently experiencing.

An exploration score can also be given individually for

each subtest area.

3. A total commitment score (the sum of all even-

numbered items) which indicates the amount of commitment

the individual is currently expressing. A commitment

score can also be given individually for each subtest

area.







63
4. Individuals can be categorized into one of Marcia's

four Identity Statuses in each content area. The test

items and their content areas are all listed following

the test in Appendix B. The scoring instructions for

assigning identity statuses are also in Appendix B.

Three methods were tested for assigning individuals

to identity statuses. A description of the methods and

the data concerning these methods are reported in the

Results section of this paper. The utility of each of

these methods is discussed in Chapter 5.

The Self-Report Measure of Exploration and Commitment

(Moberly, 1985) was also constructed for use in this study.

This instrument was designed following the suggestion

by Matteson (1977) that exploration and commitment be

measured separately using continuous scales. He rated

each of his content areas separately using a 4-point scale.

The Self-Report Measure of Exploration and Commitment

rates each content area ((1) sexual values and sex role

values; (2) dating and friendship; (3) career; and (4)

political and religious ideology) separately using a 5-point

scale. The Exploration scales range from (1) Limited

("I haven't given this area much thought") to (5) Very

Intense ("I have thought a lot about this area" or "I

am now trying to sort out my feelings about this area").

The Commitment Scales range from (1) Limited ("I don't

really have a position on this issue that I feel strongly









about") to (5) Very Intense ("I know what I think and

feel about this issue and it would be hard to change my

mind"). This scale measures the subject's overall self-

perceptions of the amount of exploration and commitment

he or she has experienced in a given area. It differs

from the Exploration and Commitment Scale in that it measures

an overall assessment while the Exploration and Commitment

Scale measures responses to specific thoughts, feelings,

and actions that may indicate exploration and commitment

in the four areas.

The Social Desirability Scale (Crowne and Marlowe,

1964) consists of 33 statements to be marked true or false.

The scale was constructed so as to strike a balance between

two types of statements; one half are culturally acceptable

but probably not true; the other half are true but not

socially desirable. The scale is designed to indicate

the effect of the social desirability motive on an indi-

vidual, i.e., the need to describe oneself in socially

desirable terms in order to gain other's approval. Of

the 33 items, 18 are keyed in the true direction and 15

are keyed in the false direction. Scores range from 0

(no social desirability) to 33 (highest social desirabil-

ity). The scale norms, drawn from 1,400 students in Intro-

ductory Psychology classes at Ohio State University, indicate

a mean for males of 15.1 (s.d. = 5.6) and for females

16.8 (s.d. = 5.5). The Social Desirability Scale has









an internal consistency (Kuder-Richardson 20) of .88,

and a test-retest correlation after a month interval of

.88. Evidence for its validity rests on many studies

in which predictive and/or concurrent hypotheses have

been confirmed. People who are susceptible to the motive

to characterize themselves in socially desirable ways

tend to (according to these studies) be more socially

conforming, to give popular word associations, to cautiously

set goals in a risk-taking venture, and to show greater

susceptibility to persuasion.

The Identity Achievement Scale (Simmons, 1970) is

a 24-item forced-choice format scale. The measure was

devised from Marcia's (1964) Ego Identity Incomplete

Sentences Blank instrument. On the Marcia instrument

sentence stems were used as stimuli. Responses to the

stems were rated by trained raters according to degree

of identity achievement indicated by the responses. Using

Marcia's (1964) scoring manual, Simmons constructed sentence

completion stems followed by two alternatives. The alterna-

tives were taken from Marcia's examples of high-, medium-,

and low-scoring responses. The alternative response

indicated by Marcia as most indicative of identity achieve-

ment was scored with a plus. Scores range from 0 to 24,

the higher scores indicating a higher level of identity

achievement. The original pool of 90 test items was reduced

to 24 through item analyses. Preliminary norms for the









scale are based on 147 freshman students in a Personality

Development class at Oregon STate University. The mean

for the sample was 12.56 (12.75 for males and 12.39 for

females). The standard deviation was 3.18. No statistically

significant sex differences in scores were found. Simmons'

measure has a test-retest Pearson product moment correlation

coefficient of .76 with 122 General Psychology students

after a one-week interval. The measure has been validated

against a number of measures including the Marcia Interview,

the Edwards Personal Preference Schedules (Edwards, 1954),

Shostrom's Personal Orientation Inventory (Shostrom, 1966),

and against aptitude test scores and grade point averages

for high school and college students (see Simmons, 1973).

In a study correlating scores on the Identity Achieve-

ment Scale with scores on the Anomie Scale (Srole, 1956),

Christie's Machiavellianism IV Scale (Christie & Geiss,

1970), Eysenck's Neuroticism's Scale (Eysenck, 1958),

Couch and Keniston (1960) Yeasaying response set scale,

Journard and Resnick (1970) Self-Disclosure Scale, and

Rotter's (1966) Locus of Control Scale, Simmons (1973)

concluded,

The tabled Pearson product-moment correla-
tions show the Identity Achievement Scale
may be assessing flexible, adaptive, func-
tioning (e.g., the negative correlation
with yeasaying and neuroticism) but is
unrelated to self-disclosing tendencies
and to perception of the locus of control
as internal. The scale seems to tap somewhat
different aspects of functioning for males
and females, e.g., it is significantly









and negatively related to the Anomie and
Mach IV scales for males but not for
females. These results suggest differential
interpretation of male and female scores.
(p. 10)


A later correlation with self-ratings of adjustment with

male and female freshman students in Psychology classes

at Oregon State University has been helpful in clarifying

the male-female differences. The pattern of specific

aspects of adjustment that correlated significantly with

the Identity Achievement Scale, "suggests that both males

and females high on Identity Achievement status emphasize

that they operate as productive and self-disciplined per-

sons. Males emphasize their lack of conformity to expected

behaviors while the females emphasize their efficiency,

social harmoniousness, clear and realistic thinking, social

approval, self-acceptance, and personal fulfillment, etc."

(Simmons, 1973, pp. 11-12).














CHAPTER IV
RESULTS


The means and standard deviations for the Exploration

and Commitment Scale are listed in Table 1. Females had

higher scores in both exploration and commitment. Both

males and females reported higher levels of commitment

than exploration. The means and standard deviations for

each of the content areas are listed in Table 5.



Table 1

Means and Standard Deviations for the Exploration and
Commitment Scale



Total Males Females
N = 245 95 150

Mean (S.D.) Mean (S.D.) Mean (S.D.)


Total scores
on Exploration
and Commitment
Scale 297.77 (28.86) 294.40 (27.38) 301.13 (30.33)

Total
Exploration 145.58 (17.93) 144.60 (17.81) 146.55 (18.08)

Total
Commitment 147.61 (20.01) 145.07 (18.49) 150.15 (21.52)









While it is clear that the various kinds of evidence

supporting the reliability and validity of a test are closely

related, I have separated the statistical results into

four areas--(1) Reliability, (2) Content Validity,

(3) Construct Validity, and (4) Concurrent Validity.


Reliability


Internal Consistency


Composite reliability (internal consistency) was com-

puted using Cronbach's alpha (1951). Three tests were

run. The internal consistency of the items making up the

Exploration Scale and of the items making up the Commitment

Scale were examined as well as the overall level of internal

consistency for the entire scale. The coefficient alpha

for the exploration items was .72; for the commitment items

the coefficient alpha was .81. The coefficient alpha for

the entire scale was .78. These alphas indicate significant

internal consistency for the two subscales as well as for

the whole Exploration and Commitment Scale.


Test-Retest Stability


The test-retest stability was explored for 45 subjects

who retook the test after a one week interval. The Pearson

product moment correlations and their associated

probabilities for the overall score, for the exploration

items and commitment items, and for the exploration and









commitment items within the four content areas can be seen

in Table 2. All except for Friendship Exploration are

significant (p < .001).


Table 2.

Test-Retest Correlations for the Exploration and
Commitment Scale


Correlation


Overall score

All exploration items

All commitment items

Career exploration items

Career commitment items

Friendship exploration items

Friendship commitment items

Political exploration items

Political commitment items

Sexual exploration items

Sexual commitment items


= 0.83397*

= 0.78976*

= 0.88861*

= 0.80181*

= 0.83511*

= 0.46162**

= 0.72607*

= 0.85017*

= 0.93087*

= 0.74093*

= 0.72268*


*p < .001. **p < .005


Content Validity


The coefficient alpha is recognized as evidence for

the content validity of an instrument. Additional evidence

is supplied by testing the following hypotheses.


Test Items









Hypothesis 1


There will be a significant positive relationship

between overall exploration scores on the Exploration and

Commitment Scale and total exploration scores on the Self-

Report Measure of Exploration and Commitment (Moberly,

1985).


Hypothesis 2


There will be a significant positive relationship

between overall commitment scores on the Exploration and

Commitment Scale and total commitment scores on the Self-

Report Measure of Exploration and Commitment (Moberly,

1985).

Both of these hypotheses were supported. Using a

Pearson product moment correlation for the comparison,

these results were obtained. Between the exploration scores

for both measures, the obtained correlation was r = .42

with a significance level of p < .001. The correlation

between the commitment items on both measures was r = .57

with a significance level of p < .001.

Further tests were conducted to examine the relationship

between the content areas on both measures. The results

can be seen in Table 3. All the correlations were signifi-

cant (p < .001)except for Career Exploration and Sexual

Exploration.









Table 3.

Content Area Correlations Between the Exploration and
Commitment Scale and the Self-Report Measure of
Exploration and Commitment


Content Areas


Correlation


Career Exploration

Career Commitment

Friendship Exploration

Friendship Commitment

Political Exploration

Political Commitment

Sexual Exploration

Sexual Commitment


*p < .001.


**p < .05.


= .07***

=.60*

=.41*

= .34*

=.52*

=.65*

= .19**

=.40*


***p > .05.


Additional evidence for the content validity of the

Exploration and Commitment Scale is provided by examining

the relationship between the Exploration and Commitment

Scale and the Simmons (1970) Identity Achievement Scale.

Two of the hypotheses are concerned with this relationship.


Hypothesis 3


There will be a significant positive relationship

between the overall scores on the Exploration and Commitment









Scale and total scores on the Simmons (1970) Identity

Achievement Scale.

This hypothesis was supported. The obtained Pearson

product moment correlation between the overall scores on

the two measures was r = .26 which is significant (p < .001).


Hypothesis 4


There will be a significant difference in mean scores

on the Simmons (1970) Identity Achievement Scale between

those who scored above the median and those who scored below

the median on the Exploration and Commitment Scale.

For this test the subjects were divided into two groups

--those who scored above the median on the Exploration and

Commitment Scale (the High group) and those who scored below

the median (the Low group). The mean score for each group

on the Simmons (1970) Identity Achievement Scale was calcu-

lated and compared. A Kruskal-Wallis test was performed

on the data producing a chi square approximation of 6.02

(p < .01) which was significant. The hypothesis was sup-

ported.


Construct Validity


Internal Consistency


The alpha levels for Internal Consistency provide some

evidence that the exploration items and the commitment items

combine to measure a coherent construct. The









convergent-divergent pattern of correlations between items

in a content area and the total content area score provides

further support for the construct validity of the Exploration

and Commitment Scale.


Convergent-Divergent Analysis


The convergent-divergent pattern of correlations between

items within a content area and their total score for the

content area on the Exploration and Commitment Scale indi-

cates that the items for each content area load moderately

to high with their own content area score while scoring

low correlations with the total scores for the other content

areas. Table 4 summarizes the range of correlations for

the individual items across the four content areas. The

ranges for the exploration items and for the commitment

items are listed separately. The median score is for the

combined range of exploration and commitment items. The

median correlations load significantly higher (p < .001)

on their own content area total than on the total scores

for the other three content areas. The items for the sexual

content area and for the interpersonal content area appear

to share some common variance. This would be expected

as interpersonal values (relating to friendship and dating)

are closely linked to sexual and sex role values.










Table 4. Convergent-Divergent Correlation Ranges for the Content Area Items
with Total Scores for Each Content Area


Total Content Area Scores


Career Ideology Interpersonal Sexual


CAREER


EXPLORATION -.00 to .63
COMMITMENT .32 to .67
MEDIAN .42*


-.03 to .22
-.06 to .20
.11


.09 to .35
-.01 to .19
.14


.02 to .32
-.02 to .16
.10


IDEOLOGY


EXPLORATION
COMMITMENT
MEDIAN

INTERPERSONAL


.02 to .13
.10 to .25
.10


.18 to .70
.34 to .73
.47*


-.10 to .34
-.05 to .21
.08


-.04 to .26
.02 to .28
.06


EXPLORATION -.04 to .28
COMMITMENT -.01 to .25
MEDIAN .17


-.01 to .23
-.04 to .18
.06


.14 to .55
.10 to .55
.39


.05 to .40
-.00 to .34
.15


SEXUAL


EXPLORATION -.19 to .17
COMMITMENT -.06 to .29
MEDIAN .14


-.08 to .15
-.00 to .18
.07


-.10 to .33
-.08 to .38
.20


.07 to .54
.11 to .53
.39


*p < .001.









Concurrent Validity


Evidence for concurrent validity is presented by com-

paring males' and females' performance in the four content

areas: (1) Career, (2) Sexual and Sex Role, (3) Inter-

personal, and (4) Political and Religious Ideology. The

data comparing the means for exploration and for commitment

for males and females can be found in Tables 5. Hypotheses

5 and 6 concern mean comparisons between males and females

in the Sexual Content area.


Hypothesis 5


The mean exploration score for females for the Sexual

Content area of the Exploration and Commitment Scale will

be higher than the mean exploration score for males in

the Sexual Content area.

The exploration mean for females was higher than for

males (females = 36.52; males = 34.27) in the Sexual Content

area. An analysis of variance (F (1, 244) = 7.61, p <

.05) showed that the difference was significant (see Table

5). The hypothesis was supported.


Hypothesis 6


The mean commitment score for females for the Sexual

Content area of the Exploration and Commitment Scale will

be higher than the mean commitment score for males in the

Sexual Content area.









The commitment mean for females was also higher than

for males (females = 40.68; males = 35.63) in the Sexual

Content area. An analysis of variance (F (1, 244) = 36.20,

E < .001) showed that the difference was significant (see

Table 5). Since the commitment means for males and females

are significantly different, this hypothesis was also

supported.

Hypotheses 8 and 9 concern comparisons between males

and females in the Interpersonal Content Area.


Hypothesis 8


The mean exploration score for females for the Inter-

personal Content area of the Exploration and Commitment

Scale will be higher than the mean exploration score for

males in the Interpersonal Content area.

The mean for females (41.91) was higher than the mean

for males (40.79); however, the difference was not signifi-

cant (F (1, 244) = 2.39, p > .05). The hypothesis was

not supported.


Hypothesis 9


The mean commitment score for females for the Inter-

personal content area of the Exploration and Commitment

Scale will be higher than the mean commitment score for

males in the Interpersonal Content area.












Table 5.

Mean Exploration and Commitment Scores (and Standard Deviations)
for Males and Females on the Exloration and Commitment Scale



Malesa Femalesb Fc


Mean (S.D.) Mean (S.D.)


Sexual Sex Role
Exploration 34.27 (5.97) 36.52 (6.36) 7.61**
Commitment 35.63 (6.61) 40.68 (6.26) 36.20***

Career
Exploration 38.65 (7.94) 39.84 (7.31) 1.43
Commitment 37.93 (10.62) 37.36 (10.89) 0.16

Interpersonal
Exploration 40.79 (5.47) 41.91 (5.53) 2.39
Commitment 38.93 (5.45) 41.18 (6.16) 8.35**

Ideology
Exploration 30.88 (8.18) 28.29 (6.82) 7.21*
Commitment 37.29 (9.02) 35.35 (7.84) 3.19

Total
Exploration 144.60 (17.81) 146.55 (18.08) 0.69

Total
Commitment 145.07 (18.49) 150.15 (21.52) 3.23

Total E & C
Score 294.40 (27.38) 301.13 (30.33) 3.08


aN = 95.

*p < .05.


N = 150.


cdf = 1.


**E < .005. ***p < .001.








The mean for females (41.18) was higher than the mean

for males (38.95). The difference was significant (F (1,244)

= 8.35, p < .005). The hypothesis was supported.

Hypothesis 11 concerns comparisons between males and

females in the Ideology content area.


Hypothesis 11


The mean exploration score for males for the Ideology

Content area of the Exploration and Commitment Scale will

be higher than the mean exploration score for females.

The male exploration mean (30.88) was higher than

the female mean (28.29). The difference was significant

(F (1,244) = 7.21, p < .05). The hypothesis was supported.

There were no hypotheses regarding differences in

means in the career area for males and females because

results have been mixed in this area. It is interesting

to note that the means for Career Commitment show the small-

est difference of any of the areas (males = 37.93, females

= 37.36) (F (1,244) = .16, p > .1).

In summary, it appears that females generally had

higher levels of exploration and commitment. Significant

differences were found in four areas. Females had signifi-

cantly higher levels of sexual exploration, sexual commit-

ment, and interpersonalcommitment. Males had a significantly

higher level of exploration in the Ideology area. The

smallest difference between males and females was found









in the area of career commitment suggesting that females

are as committed to career goals as males are. Interpersonal

exploration and ideological commitment appear to be important

to both males and females.

There were two hypotheses that were related to comparing

males and females within Marcia's Identity Statuses.


Hypothesis 7


Females will be significantly more likely than males

to be categorized in the more advanced ego identity statuses

(Achievement and Moratorium) in the Sexual Content area.


Hypothesis 10


Females will be significantly more likely than males

to be categorized in the more advanced ego identity statuses

(Achievement and Moratorium) in the Interpersonal Content

area.

Three methods of placing subjects in Marcia's Identity

Statuses were explored. A "ranking" method was tried in

which exploration and commitment scores were rank ordered.

The top third were considered high and the bottom third

low. Achievements were high on both exploration and commit-

ment. The diffusions were low on both dimensions. Mora-

toriums were high in exploration and low in commitment.

Foreclosures were low in exploration and high in commitment.

The subjects in the middle third wre not placed in statuses.









The "mean-split" method was also explored. Those

above the mean were considered high; those below the mean

were considered low. Achievements were high in both explora-

tion and commitment. Diffusions were low in both dimen-

sions. Foreclosures were high in commitment and low in

exploration. Moratoriums were high in exploration and low

in commitment.

The third method--the "standard deviation" method--

established a mean and a standard deviation for each total

content area score. Those who scored one standard deviation

above the mean were Achievements. Those who scored one

standard deviation below the mean were Diffusions. The

subjects who scored between the two standard deviations

were classed as either Foreclosures or Moratoriums depending

on how their Exploration score compared to their Commitment

score. If the commitment score was higher by 10 points,

they were considered Foreclosures. Otherwise, they were

considered Moratoriums.

A comparison of the percentages of males and females

in the more advanced identity statuses vs. the low identity

statuses for the various content areas was computed for

all three methods using the chi square statistic. Results

are summarized in Tables 6, 7, and 8. There were no sig-

nificant differences for any of the content areas.

As can be seen from Tables 6, 7, and 8, hypotheses

7 and 10 were not supported. There were no significant













Table 6.


Standard Deviation Method. Male-Female Differences in Identity
Status Classification Based on Content Areas of the Exploration
and Commitment Scale


Chi
Content Area Identity Statuses Squarea


High Low


Career


Males


63%


Females 59%


Sexual and
Sex Role
Values

Political
and Religious
Ideology

Interpersonal
Area (Friend-
ship and
Dating)


Males

Females

Males

Females

Males

Females


Males N = 95. Females N =

adf = 1


67%

58%

58%


54.7%


37%

41%

33%

42%

42%


45.3%


73%

67%


27%

33%


.34*


2.47*


.25*


.77*


150.


*p > .10













Table 7.

Mean Split Method. Male-Female Differences in Identity Status
Classification Based on Content Areas of the Exploration and
Commitment Scale


Chi
Content Identity Statuses Squarea


High Low


Career Males 50.5% 49.5%
.34*
Females 52% 48%


Sexual and Males 47.4% 54.7
Sex Role 1.24*
Values Females 54.7% 45.3%


Political Males 47.3% 52.7%
and Religious .86*
Ideology Females 41.3% 58.7%


Interpersonal Males 58% 42%
Area (Friend- .35*
ship and Females 54% 46%
Dating)


Males N = 95. Females N =

adf = 1.


150.


*p > .10.













Table 8.

Ranking Method. Male-Female Differences in Identity Status
Classification Based on Content Areas of the Exploration
and Commitment Scale


Chi
Content Area Identity Statuses Squarea


High Low N


Career Males 41.46% 58.54% 54
1.79*
Females 55.00% 45.00% 90
Total 144
Sexual and Males 34.88% 65.12% 52
Sex Role 3.74***
Values Females 53.62% 46.38% 81
Total 133
Political Males 63.27% 36.73% 46
and Religious 5.90**
Ideology Females 40.58% 59.42% 81
Total 127
Interpersonal Males 43.18% 56.82% 51
Area (Friend- 1.18*
ship and Females 53.73% 46.27% 83
Dating) Total 134


Males N = 95. Females N = 150.

df = 1


*p > .10. ** K < .05.


***p > .05.









differences in the percentages of males and females in

the more advanced ego identity statuses in the Sexual and

Interpersonal Content areas.

A post hoc analysis was conducted in the four content

areas. Males and females were compared in the Committed

(Achievement and Foreclosure) Statuses vs. the Uncommitted

(Moratorium and Diffusion) Statuses, using the Standard

Deviation method of placing subjects in statuses. Results

are seen in Table 9.

The only significant difference was found in the area

of Sexual and Sex Role Values. Females were significantly

more likely to be found in the Committed Statuses in this

area.

Table 10 contains the numbers and percentages of males

and females in each identity status as classified by the

Standard Deviation Method.


Social Desirability Bias


The question of whether the test responses were

contaminated by a social desirability bias was examined

by testing the following hypothesis.


Hypothesis 10


There will be no significant relationship between

the overall scores on the Exploration and Commitment Scale













Table 9.

Male-Female Differences in Committed v. Uncommitted Statuses Based
on Content Areas of the Exploration and Commitment Scale


Chi
Content Area Identity Statuses Squarea


Committed Uncommitted


Career Males 41% 59%
.23**
Females 38% 62%


Sexual and Males 30% 70%
Sex Role 5.33*
Values Females 46% 54%


Political Males 43% 57%
and Religious .15**
Ideology Females 40.6% 59.4%


Interpersonal Males 34% 66%
Area (Friend- .0032**
ship and Females 33% 67%
Dating)


Males N = 95.


Females N = 150


adf = 1


*p < .05. **p > .1.









Table 10.

Numbers and Percentages of Males and Females in Each Identity
Status as Classified by the Standard Deviation Method



Males Females


Number Percentage Number Percentage


Career Content Area

Achievement 18 19% 30 20%
Moratorium 38 40% 64 43%
Foreclosure 21 22% 27 18%
Diffusion 18 19% 29 19%

Interpersonal Content Area

Achievement 30 32% 41 27%
Moratorium 39 41% 60 40%
Foreclosure 2 2% 9 6%
Diffusion 24 25% 40 27%

Sexual Content Area

Achievement 20 21% 37 25%
Moratorium 44 46% 49 33%
Foreclosure 9 9% 31 21%
Diffusion 22 24% 33 21%

Ideology Content Area

Achievement 26 27% 38 25.3%
Moratorium 29 31% 44 29.4%
Foreclosure 15 16% 23 15.3%
Diffusion 25 26% 45 30%


Males N = 95. Females N = 150.







88

and the total scores on the Crowne and Marlowe (1964) Social

Desirability Scale.

This hypothesis was supported. A Pearson product

moment correlation (r = .11) was not significant (p > .05).


Summary


This study provides evidence for the adequacy of the

reliability, content, and construct validity of the Explora-

tion and Commitment Scale. The comparison of means data

were consistent with previous research results, providing

evidence of concurrent validity for the scale. Social

desirability bias does not appear to affect subjects'

responses to the test items. The hypotheses relating to

classifying subjects into Marcia's statuses were not

supported. Possible reasons for these results will be

discussed in Chapter 5. Finding a significantly greater

percentage of females in the committed statuses in the

Sexual Content Area was consistent with previous research

findings. This result will also be discussed in Chapter

5.














CHAPTER V
DISCUSSION


This study provides some evidence for the reliability

and validity of the Exploration and Commitment Scale. The

scale's internal consistency, test-retest reliability,

the convergence and divergence of subscale items, and its

content, construct, and concurrent validity have all been

investigated.


Reliability Data


The two measures of reliability (test-retest and inter-

nal consistency) were both significantly supported, showing

the test to be a reliable measure. Internal consistency

results will be discussed in conjunction with the conver-

gence-divergence data in the Construct Validity Section.


Content Validity


Correlation with Measurement of Ego Identity


It is important in validating an instrument to provide

some empirical evidence that the scale measures the construct

you intend to measure. The significant correlation with

the Simmons Identity Achievement Scale (IAS) (1970) is

a first step in providing empirical evidence of the content

89









validity of the Exploration and Commitment Scale. The

IAS is based on Marcia's (1964) Ego Identity Incomplete

Sentence Blank, against which he validated his Identity

Status Interview. Evidence for the validity of the IAS

can be found in the instrument description in Chapter 3.

The IAS measures an overall sense of identity development

without discriminating as to the mode of identity resolu-

tion. Simmons does not suggest that the IAS will place

subjects in an identity status.

The two tests that compared the Exploration and Commit-

ment Scale to the IAS were both significant. They provided

two different kinds of evidence for content validity.

The overall correlation shows that the subjects who scored

high on the Exploration and Commitment Scale scored high

on the IAS. The comparison of means demonstrates a signifi-

cant difference between high and low scorers on each meas-

ure. Not only does the Exploration and Commitment Scale

correlate with an instrument that measures ego identity,

but both instruments discriminate significantly among a

range of scores.


Correlation with Self-Report Measure


While the above data support the contention that the

test measures ego identity as it is theoretically described,

the correlations with the Self-Report Measure of Exploration

and Commitment (Moberly, 1985) provide another kind of









evidence. The self-report measure was designed to ask

directly for an individual's overall sense of the amount

of exploration and commitment he or she has experienced

in each of the four areas. The Exploration and Commitment

Scale asks for responses to specific beliefs, behaviors,

and experiences that less directly indicate exploration

or commitment. That the exploration items in the two meas-

ures and the commitment items in each of the measures cor-

relate significantly suggests that the test items do reflect

the individual's experience of exploration and commitment

and thus do indeed sample the content domain of ego

identity.


Construct Validity


Internal Consistency and Convergence-Divergence Data


The construct validity of the instrument is supported

both by its level of internal consistency and by the conver-

gence-divergence data. That the levels of internal consis-

tency were acceptably high not only for the two subsections

of exploration items and commitment items but for the overall

test is important. When this information is combined with

the convergence-divergence data, one gets a sense for the

flexibility of the instrument as a measurement of the con-

struct, ego identity. The converence-divergence data show

that the items in the four content area subscales (Sexual,

Interpersonal, Ideology, and Career) load moderately high








on their own subscale totals while having low correlations

with the other subscale totals. This supports the belief

that the four content area subscales measure discriminately

different areas of identity development.

The construct, ego identity, as it has been discussed

by Erik Erikson (1959, 1968) and James Marcia (1964, 1966),

combines the two variables of exploration and commitment.

Both are closely linked to ego identity development. The

ego expands through exploring options and testing out pos-

sibilities. One's sense of personhood is enlarged as the

individual is able to try himself or herself out through

new experiences. A sense of one's similarity to others

is important, but so is the ability to savour the differences

between oneself and others. A sense of individuality is

built on both.

As the individual explores, he or she chooses values,

goals, or options that provide self-definition. It is

from the vantage point of one's commitments that the indi-

vidual is able to see the similarities or differences between

himself or herself and others. In making commitments that

one will stand by, the individual is giving up possibili-

ties. Deciding to be a lawyer or a Methodist or monogamous

means that one has chosen to limit oneself to a particular

career, religion, or sexual status. The goal in individua-

tion could best be described as finding a comfortable balance

between these two variables. Exploring enough so that




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs