0.: L: vow.... £31.... 1.: WWWW WWW WWWW WWWW WWW LIB7RXK I Michigan State University This is to certify that the dissertation entitled Ego—Identity Development of Early Adolescents presented by Mark Picciotto has been accepted towards fulfillment of the requirements for Eh 0 degree in CounsehngJ' sychol ogy WWW /2/ Major professor / Date—2/6/87 “(Ilka— Ak' 4‘ l ’ "‘1 lnrr - 1 - . 0.1 1 MSU RETURNING MATERIALS: Place in book drop to LWBRARJES remove this checkout from .—:—. your record. FINES will be charged if book is returned after the date stamped below. 0 q 2 .t {u "a “(in H7- ”2.513 5 EGO-IDENTITY DEVELOPMENT OF EARLY ADOLESCENTS Mark Picciotto A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY School of Health Education Counseling Psychology and Human Performance 1987 ABSTRACT EGO-IDENTITY DEVELOPMENT OF EARLY ADOLESCENTS By Mark Picciotto Part of the reason there has been a relative paucity of research on the identity development of younger adolescents is the lack of valid objective measures of identity status. Further, there is a widely held belief that the important changes in identity status do not occur until later adolescence, in spite of a small but significant body of research which disputes this assumption. In the present study, the Objective Measure of Ego Identity Status (ORIEIS), and a shortened version of the Assessment of Adult Adjustment Patterns (AAAP) were administered to 214 students between the ages of 13 and 18, in the 8th, 10th, and 12th grades. The purpose of the study was to identify any changes in identity status with increasing age, and to attempt to partially validate the AAAP on the ONIEIS. Although a high number of subjects (about 50%) were not classifiable in any of the identity statuses or appeared not to have mastered any Eriksonian stages, there was a significant shift with age both out of the foreclosure status, and into the achievement status which was correlated with social desirability. The only significant relationship between the OMEIS and the AAAP was a canonical correlation between them. However, a Guttman scale analysis of the AAAP, which is presumed to be unidimensional and hierarchical, showed it to have quite good qualities as a Guttman scale. Also, a MANOVA analysis of the theoretical underlying variables of the OMEIS, the tendency to consider life alternatives and a commitment to one of them, showed a significant increase with age in the tendency to consider alternatives, but not in the tendency to commit to one of them. Thus, the conclusions were drawn that significant identity development does occur in younger adolescence, that both the Oh/IEIS and the AAAP appear promising for use with this age group, but that either the cutoff levels of both tests should be revised downward, or future researchers should directly use the interval level data which the tests produce, not the mastery concept. ACKNOWLEDGEMENTS I wish to thank all the members of my committee, but most particularly Dr. William Farquhar who, as chairman, provided invaluable support and assistance throughout the entire process of completing my dissertation. Dr. Steven Raudenbusch provided very valuable suggestions on questions of analysis and interpretation of results. Dr. Richard Houang was available and helpful throughout the time I was working, and Dr. Alton Kirk was supportive and most flexible. In the Lansing area schools, I wish to thank Mr. Anthony Waldron and Mr. Paul Berdon, principals of the high school and middle school respectively, where the data was collected, for their willing accomodation to my needs. I also wish to thank the many teachers who helped, as well as the students who took the time to fill out the questionnaires. Other people without whom this dissertation would not have gotten completed include Anthony Straseske and Lyn Gay, who were very helpful during the data collection process, as well as being willing to run errands during all phases of the project, and allowing their house to be used as a base. Jeffrey Picciotto has been willing to take many hours of extra time to make sure that the printed version of the dissertation was properly edited and presented. My wife Yoshie has been helpful in more ways than I can mention providing support, typing, drafting, and love. Thank you all. iv Chapter [0 TABLE OF CONTENTS LIST OF ILLUSTRATIONS LIST OF ABBREVIATIONS THE PROBLEM PURPOSE NEED OVERVIEW OF THE THEORY Erickson’s Theory of Development Marcia’s Four Identity Statuses STATEMENT OF THE HYPOTI-IESES IN BROAD RESEARCH FORM OVERVIEW OF THE THESIS REVIEW OF LITERATURE STUDIES USING MEASURES OTHER THAN THE MARCIA INTERVIEW OR THE OMEIS STUDIES USING THE MARCIA INTERVIEW TO ASSESS IDENTITY STATUS STUDIES USING THE OBJECTIVE MEASURE OF EGO IDENTITY STATUS (OMEIS) METHODOLOGY RESULTS SAMPLE AND POPULATION INSTRUMENTS Fact Sheet Objective measure of Ego Identity Status (OhIEIS) Assessment of Adult Adjustment Patterns (AAAP) PILOT STUDY ADMINISTRATION OF THE MATERLALS DESIGN OF THE STUDY HYPOTHESES IN TESTABLE FORM ANALYSIS SUMMARY DISTRIBUTION OF CASES ON THE OBJECTIVE MEASURE OF EGO IDENTITY STATUS Page viii Auto 0—- CDA 13 14 15 15 21 25 32 33 33 33 36 38 39 39 4O 43 47 48 48 TABLE OF CONTENTS (Concluded) Chapter Page DISTRIBUTION OF CASES ON TIIE ASSESSIV'IENT OF ADULT ADJUSTMENT PATTERNS 49 EFFECT OF THE CORRECTION FOR SOCIAL DESIRABILITY 50 TESTS OF THE HYPOTHESES 53 5 SUINIMARY AND CONCLUSIONS 66 REVIEW OF TIIE THEORY 66 REVIEW OF RESEARCH 67 REVIEW OF hIETHODOLOGY 68 REVIEW OF RESULTS AND CONCLUSIONS 69 DISCUSSION 76 LIMITATIONS OF THE STUDY 79 SUGGESTIONS FOR FUTURE RESEARCH 80 APPENDIX A 81 APPENDIX B 8‘2 APPENDIX C - 83 APPENDIX D 84 APPENDIX E 87 BIBLIOGRAPHY 100 vi Figure pk 00 IO 0‘ ‘3 10 ll 13 14 LIST OF ILLUSTRATIONS The four identity statuses of Marcia’s theory of development Patterns of identity development according to Waterman (1982) Summary of Research on the Identity Status of Early Adolescents Descriptive statistics over the whole sample (n=214) for the OMEIS Pearson correlation coefficients of the identity statuses of the OlVIEIS over the whole sample (n=214) Number of students in each identity status by grade level as classified by scores on the OhIEIS Number of students in each Eriksonian stage by grade level, classified by mastery level at each stage Reliabilities of the OMEIS and AAAP Number of cases in each identity status by grade level, when cases are eliminated for responding in a highly socially desirable way Number of cases in each developmental stage by grade level when cases are eliminated for responding in a highly socially desirable way Number of students failing stage 1, ‘2, 3, or 4 who are in diffusion Number of students failing stage 1, 2, 3, or 4 who are in diffusion with exclusion for high social desirability in effect Number of students who have mastered stages 1, 2, 3, and 4 who are in moratorium Number of students who have mastered stages 1, 2, 3, and 4 who are in moratorium, with exclusion for high social desirability Number of 12th grade students in moratorium who have mastered stage 4 Number of 12th grade students in moratorium who have mastered stage 4, with social desirability exclusion in effect Coefficients of the two sets of canonical variates Pearson correlations between the identity statuses of the OhIEIS and the four subfactors of Stage 5 of the AAAP Number of all students in diffusion who have failed to master stages 1, ‘2, 3. or 4 on the AAAP, using the new cutoff levels Number of all students in moratorium who have mastered stages 1, 2, 3, and 4, on the AAAP, using the new cutoff points Number of 12th graders in moratorium who have mastered stages 1, 2, 3. and 4, on the AAAP, using the new cutoff points Graph of average identity status scores for each grade level Graph of average transformed variable scores by grade level vii Page 10 ll 48 49 0| *l U! ~l o: (I) 59 60 60 61 AAAP : OMEIS: A or ACH: D or DIF: F or FOR: M or MOR: LIST OF ABBREVIATIONS Assessment of Adult Adjustment Patterns Objective Measure of Ego Identity Status Identity Achievement Identity Difl'usion Identity Foreclosure Moratorium viii CHAPTER 1 TIIE PROBLEM Answers to the questions ”\Vho am I?” and ”How did I become this way?” have been proposed since antiquity. Only relatively recently however, have complete, fairly well integrated theories arisen to address the issue of development in general, and the development of a sense of identity in particular. One of the first was Freud’s psychoanalytic theory, which strongly emphasized a child’s early life, and maintains that an individual’s identity is determined by a complex system of internal adjustments. Thus, the important components of an individual’s identity and of identity development are within an individual’s psyche, and are influenced chiefly by the parents. Moreover, an individual’s identity is set at a fairly young age, and is relatively resistant to change. There are several variants to the psychoanalytic approach. A second general theory is the behavioral one, which in its most extreme form maintains that an individual has no personality per se, but rather is completely shaped by a series of positive and negative reinforcers to behaviors. A person then becomes essentially a ”reaction machine” who behaves in every situation the way predicted by previous reinforcers. In this system, the environment plays a paramount role in the determination of an individual’s identity. There are also several variants to the behavioral approach. A third broad class of theories are those which attempt to combine a sense of the importance of the individual’s inner world with an appreciation of the importance of the individual’s environment as well. One of the most important of these has been Erik Erikson’s theory of development (Erikson, 1956, 1968). As a psycho-social theory, it recognizes that an individual’s inner development is important, but maintains that a sense of identity is equally influenced by the social matrix within which one must define one’s identity. Erikson’s developmental theory was one of the first which truly recognized development as a life-long process, and moreover, considered a sense of identity to be one of the central organizing concepts of that process. As theories which purported to explain the development of a sense of identity emerged, so did research which attempted to confirm or deny the truth of each theory. Erikson's theory in particular, has given rise to a great deal of research because of the integrating features present in it not present in other theories. Unfortunately, Erikson’s postulates are written in a nonspecific way, making them particularly difficult to operationalize. The concept of identity although central to the theory, was resistant to empirical testing. In 1964, James Marcia developed a semi-structured interview to assess an individual’s identity status, a specification of the type of identity one had resolved, derived from his interpretation of Erikson’s writing (Marcia, 1964). Marcia’s operationalization of Erikson’s concepts proved useful, but the interview was cumbersome, took a lot of time to administer, and required a great deal of time to train interviewers who had high reliability. In response to this need, a number of more objective tests of identity status were produced, one of the most useful of which was one developed by Adams (Adams, Shea, and Fitch, 1979). It was tested and partially validated on college students. Independently, Farquhar, Parmeter, and Wilson developed a test (the Assessment of Adult Adjustment Patterns, AAAP) to determine one’s stage in the Eriksonian development process which has had few validity studies (Farquhar, Wilson, and Parmeter, 1977). PURPOSE The purpose of the present study was two fold. On the one hand the usefulness of the developmental scheme developed by Marcia (see Marcia’s Four Identity Statuses, this chapter), was investigated on a population younger than that upon which it was originally validated. In the study not only were the constructs of identity diffusion, foreclosure, moratorium, and identity achievement examined for their expected distributions in this population, but the presumed underlying senses of developing crisis and commitment were also investigated. The second part of the study was a validation of a portion of the AAAP. It arose out of the fact that both the AAAP and the ONIEIS were developed with direct reference to Erikson’s theory of development, and therefore were presumably measuring through different instruments the same underlying developmental process. For example, an individual who appeared in the ’identity achievement’ status (the highest) of the OhIEIS, was expected to have fully mastered the Stage 5 (Identity) portion of the AAAP, since that individual would be near the end of his identity development tasks. NEED The need for this study derives from the fact that a great deal of the research to date in identity development has been done with college students, because of the convenience in obtaining samples. This is true in spite of the fact that Erikson was far from specific about the ages at which he presumed this adolescent crisis to occur. It is possible that many of the changes described by Erikson could occur, or begin to occur, in younger adolescents than those of college age. However, a strikingly small number of studies have investigated this age group. This study helps to fill that gap. Another difficulty in doing research in the field of identity development is the lack of valid measures. This study, in moving closer to a validation of both instruments being used, helps in that regard. A final area of need addressed by this study is the fact that one of the most readily accepted instruments in identity development research, the Marcia semi-structured interview, is also one of the most time consuming to administer and difficult to score. The instrument developed by Adams has been concurrently validated with the semi-structured interview. The additional validation of this paper and pencil test of identity status is helpful to future research by making the test more objective, easier to score, and allowing larger samples to be tested. OVERVIEW OF THE TIIEORY The theory being used for this study is Erik Erikson’s theory of human development, with James Marcia’s redefinitions of the stage of identity development. In this section, the essential aspects of Erikson’s theory will be described, as will the ways Marcia has made Erikson’s identity stage more precise. Finally, the way this study relates to Erikson’s theory will be clarified. Erikson’s Theory ofDevelopment Erikson’s original interest was in abnormal rather than normal identity formation. Noting the difficulties which some World War II veterans had in reentering society, he labeled those problems ”acute identity diffusion”. Erikson came to believe that the problems the veterans had in leaving one role and entering another were psychologically similar to the problems which some adolescents have in leaving childhood and entering adulthood. His work led him to construct a theory of development, with identity as its main anchoring concept, which runs from infancy to old age, encompassing eight distinct stages. The first four stages involve the development of aspects of the self which are felt to be crucial precursors to identity formation, the fifth stage involves the development of one’s sense of identity, and the last three deal with the evolution of identity as the individual progresses through life. Before elaborating on the stages in a more precise manner, several aspects of the theory need to be made clear. First, it should be acknowledged that, as a trained psychoanalyst himself, Erikson did not underestimate the role played by processes internal to the self in identity formation. In fact he postulated two types of identity: ego identity which refers to one’s commitment to social roles such as work, religious values or political beliefs. Self identity refers to the individual’s self-perceptions and role-images. Although the development of the two overlap to a fair extent, Erikson considered self-identity to be a primary focus at younger ages, and ego-identity to be developed more intensively during adolescence and young adulthood. Most studies of identity development in adolescence deal implicitly or explicitly with ego- identity, as does this study. Second, Erikson viewed development as a series of tasks which must be accomplished by the individual. Thus, each stage is defined as a ”crisis”, which must be overcome or solved in the appropriate way, or ones development lapses into a maladaptive state. For this reason, each of Erikson’s stages is described as a choice of alternatives, one good and one bad, which describe the adaptive and maladaptive states of the resolution of a stage. For example, the second stage of the theory is called ”autonomy vs shame and doubt”, reflecting a sense of autonomy as a desirable outcome, with a sense of self-doubt as the maladaptive alternative. Third, Erikson’s theory, as are most stage theories, is cumulative. That is, an individual is expected to use the resolution of all previous stages as a basis for the resolution of the next crisis. It is presumed that inadequate or poor resolution of a developmental task seriously, if not totally impairs further development, since building blocks needed for further development have not been provided. The eight stages of human development are briefly described in the following section. Stage 1: Trust vs. Mistrust In Stage 1, the infant from birth to about one year gains both a sense of trustfulness of others as well as a sense of his own trustworthiness. The physical basis of the development is the process of ”taking in” both sustenance and various stimuli, and growing appreciation of taking in, and later, of giving back. The feeling of trust in identity arises out of the encounter between mother and infant, which is one of mutual trustworthiness and recognition. Stage 1 is also the basis for the capacity for faith, which is echoed in the importance of the institution of religion. Stage 2: Autonomy vs. Shame and Doubt Stage 2 is a time for the child to develop a sense of separateness from the parents, particularly the mother. The separateness entails both an expression of one’s personal will, and a growing understanding that one is, to a certain extent, controlled by the parents. The psychological manifestation of one’s will is through a propensity to ”hold on” to what is one’s own, or to release it. The contribution made to a sense of identity from Stage 2 is the gaining of the courage to turn away from others to be one’s self, and the danger is that the self-doubt that arises upon parental control that is felt as too strong impairs one’s will to self-determine. The human institution echoed from Stage 2 is that of law and order, which is supposed to control, yet allow one the freedom to be one’s self. Stage 3: Initiative vs. Guilt In Stage 3, which takes place around the age of three, the child’s increasing ability to speak and to move freely about allows him to greatly expand his horizons and his expectations about what he can do to fulfill his capacities. At age three, a growing awareness of sexual differences, alliance with the opposite sex parent, and a growing sense of conscience, gives rise to a sense of guilt. The contribution of Stage 3 to a sense of identity lies in freeing the imagination and leaving in place the initiative to attempt to become all one can. Stage 4: Industry vs. Inferiority Stage 4, which coincides with the beginning of school, marks the beginning of the child’s identification with the world of work as well as that of play. The child takes pride in being a ”maker” as well as a player, and can try on and discard many potential roles from the adult world of work. The pleasure of doing tasks well and of being recognized for them fosters a sense of industry, while an estrangement from what one can accomplish may give rise to a sense of one’s inferiority, particularly as social connections, peer groups, and teachers become much more important to one’s development. Stage 5: Identity vs. Role Confusion Stage 5, which begins with the onset of puberty, is the stage in which previous psychosocial crises are reexamined in light of the now pressing need to integrate them into a more stable, ”final” identity. Thus, it is important to have a moratorium, or a period during which an integration can take place, and during which each earlier stage becomes reworked in the service of an adult, integrated identity. The need for trust manifests itself as a need to have idols and ideas to have faith in, as well as ones in whose service one can prove one’s self trustworthy. A sense of one’s will makes it important that the adolescent decide freely upon one of the available avenues for self- fulfillment, and will oppose being forced into activities that would expose him to shame or self- doubt. The legacy of the third stage is the willingness of the adolescent to allow his imagination to expand the range of possible avenues of fulfillment. The desire to make things work well makes the choice of occupation important beyond questions of money and status, and will sometimes make one refuse a job in which one would be producing in a manner inferior to one’s capabilities. The danger in Stage 5 is identity confusion, a state in which one is unable to adequately define one’s self as a meaningful individual in society. The dilemma of choosing a profession is often the most disturbing element of identity confusion, yet it can manifest itself in an inability to define one’s values and meanings in a variety of social arenas, such as the religious, ethnic, cultural, and political. Stages 6, 7, and 8: Beyond Identity Following the establishment of a sense of identity, one may develop a full sense of intimacy with another, that is, the ability to fuse one’s self with another while still retaining one’s core identity. The counter to the full ability to share of one’s self is a sense of isolation reflected in an unwillingness to risk an undeveloped or fragile identity by exposing one’s self to another. The next psychosocial crisis involves the establishment and guiding of the next generation. The development of Stage 7 is a more mature form of expression of man’s need to be needed, as well as his needs for productivity and creativity. The sense of creativity need not express itself only in child-rearing, but where it is not present at all, a sense of stagnation arises, in which one may begin to indulge one’s self as if one were a child, yet without an accompanying sense of satisfaction. Finally, the aging person who has accepted her life cycle for what it has been, has accepted her responsibilities to herself and to others, and is able to make sense of her life both for herself and in its larger social context, has achieved a state of ego integrity. Failing to achieve ego integrity, one might be left with a sense of despair that life is too short, that one would like to, but cannot, start life anew and try other roads. One has a sense of disgust at life in general, reflecting one’s disgust of one’s self. Marcia ’3 Four Identity Statuses Faced with a rich theory that was nevertheless difl'icult to test, several theorists stepped forward with clarifications and expansions of Erikson’s theory. In the area of identity development, however, none has been so influential as Marcia (1966). He proposed four different ways of resolving the identity crisis, each of which he called an identity status. Moreover, each identity status could be defined by the presence or absence of two crucial ingredients in an adolescent’s life: an active consideration of several alternatives in a given psychosocial realm, and a clear commitment to one of them. Thus one could have neither considered alternatives, nor committed one’s self to any, and one would be in the identity diffusion status. If one were actively considering options, without having chosen any yet, one would be in the moratorium status suggested by Erikson. If one had considered a variety of options, and had committed to one of them, one would be in the identity achievement status for that psychosocial realm, for example, occupation. Finally, if one had committed one‘s self to an option without having considered the possible alternatives, one would be considered identity foreclosed. Identity foreclosure might happen when a parent’s values are accepted by the adolescent without questioning. In Figure 1, an illustration of the four identity statuses is provided. It is clear from both Erikson and Marcia that the ideal path to identity achievement lies in starting out identity diffused, as anyone is at a young enough age, going into a period of moratorium, or active consideration of one’s choices, then settling on those that make sense for one, that is, becoming identity achieved. However, other possible paths exist, for example an individual may become identity foreclosed for a time, then, after becoming aware of alternatives, Commitment absence presence M = moratorium presence M A A = identity achievement Crisis D = identity diffusion absence D F F = foreclosure Figure 1. The four identity statuses of Marcia’s theory of development begin considering them (moratorium), then choose one more congruent with her needs (identity achievement). Furthermore, Erikson’s theory makes it quite allowable to ”regress” from a higher status to a lower one in the short term, so as to achieve a more fulfilling long term adjustment. Regression would occur if an individual in the process of considering alternatives (moratorium). decided to simply postpone such debate and return to a state of identity diffusion, either because no satisfactory resolution seemed to exist, or because more pressing concerns emerged. Waterman (1982) has proposed a model of the sequential patterns of ego development. The model suggests a developmental pathway of development, illustrated in Figure 2. As can be seen, not all mathematically possible pathways are considered theoretically possible. For example, individuals in either moratorium or identity achievement cannot become identity foreclosed, because both the moratorium and identity achievement statuses involve the contemplation of alternatives, while foreclosure indicates commitment without considering alternatives. For a similar reason, an individual in identity foreclosure cannot directly pass into identity achievement, since it would not seem possible to pass from a state of being committed without having considered alternatives to one of being committed to one of several alternatives, without first passing through a period of questioning, or moratorium. 10 START STAGE 5 IDENTITY DIFFUSION In there cnsrs? roncwsune “6—! HOIATDRIUH COHHITHENT atned? IDENTITY ACHIEVEMENT YES PROCEED TO STAGE 6 Figure 2. Patterns of identity development according to Waterman (1982) 11 Drawing upon Erikson’s work as well as psychoanalytic theory, ego psychoanalytic theory, and cognitive-structural developmental theory, Marcia (1983) has recently proposed three related variables which, having been measured at early adolescence, should allow one to predict identity resolution at the end of adolescence. The three variables that are proposed are rooted in ego-developmental theory. Marcia’s reasoning goes that as an adolescent has to cope with the increasing disorder of his or her life, there is a need for controls. The control most used by children is the superego or conscience. The problem is that the injunctions of right and wrong contained in the childhood superego are not appropriate to adolescents, so it is increasingly ignored and the adolescent begins to enlarge upon the already partially formed ego-ideal, to slowly make it into that which is peculiarly right for him or her. \Vhen identity development is viewed in this light, Marcia states that three variables will be of hierarchical importance: confidence in parental support, a sense of industry, and a self-reflective approach to one’s future. Confidence in parental support is a prerequisite to allowing an adolescent to relinquish superego control, since temporary external control will be needed, and the adolescent must be able to expect it from the parents. Such confidence is essentially a restatement of Erikson’s formulation that ”basic trust” in the parents is necessary for the adolescent to embark on identity exploration. A sense of industry is critical for the child to feel competent in his or her skills and feel that mastery of a given area is possible. An additional advantage is that competence and mastery lead to high self esteem. From a theoretical point of view, mastery of the last psychosocial task before identity should be a good predictor of mastery of the identity task. 12 A self reflective approach to the future implies an ability to see one’s self in the future, to introspect about one’s self and one’s abilities. Such introspection implies formal operations, but includes as well an accurate knowledge of who one is, what one ”has to offer” and some possible avenues for expression of interests and talents. Marcia hypothesizes that the three variables are hierarchical, with confidence in parental support at the base, followed by industry and self-reflection. The reasons are largely based on Erikson’s theories: one must have trust before one can work on industry, and the first two are necessary to work on the issues directly pertaining to identity. In the preceding ways, then, Erikson’s theory of human development, in which the development of a sense of identity is the central concept, has been made more precise and theoretically expanded by the efforts of Marcia and Waterman. STATEMENT OF THE HYPOTIIESES IN BROAD RESEARCH FORM The following hypotheses relate to the validation of the OMEIS. 1. The proportion of students at each grade level in the lowest level of identity development «identity diffusion-- decreases as grade level increases (8th to 10th to 12th grade). 2. The proportion of students at each grade level in the most active phase of identity development amoratorium— increases as grade level increases. 3. The proportion of students at each grade level in a phase of identity development characterized by premature termination of identity development ~foreclosure~ does not change with grade level. 13 4. The proportion of students at each grade level who are identity achieved remains constant and low across grade levels. The following hypotheses relate to the validation of the constructs of the AAAP. 1. Any student who has an unresolved developmental task prior to identity development, is in the lowest identity status. To Any student who has no unresolved early developmental task is in moratorium. 3. Any high school senior who has resolved the developmental task immediately prior to identity development is in moratorium. 4. Each of the identity statuses on the OMEIS corresponds to one of the subfactors of the identity crises portion of the AAAP. OVERVIEW OF THESIS The remainder of the research is presented as follows: in Chapter 2 the relevant research pertaining to the development of a sense of identity in early adolescents is described, in Chapter 3 the measures used in this study, the methodology and the types of analysis of the data are presented, in Chapter 4 the results of the data analysis and answers to the research questions are presented, and in Chapter 5 the thesis is summarized and possible avenues for future research are raised. 14 CHAPTER 2 REVEW OF LITERATURE The great bulk of research in identity development has been concerned with older adolescents, particularly those of college age. There have thus been only a few studies pertaining to younger adolescents, and they are reviewed in this chapter. For the purposes of the review, a younger adolescent is one who is between 13 and 18 years of age, or is in middle school or high school. Older adolescents are those in college or older than 18 years old. Since the current study describes the relationship between two measures of identity development, the review is organized by the nature of the measure used to assess identity. In the first section studies using measures other than Marcia’s Ego-Identity Status Interview or Adam’s Objective Measure of Ego—Identity Status (OMEIS) are described. In the second section studies which have used the interview to assess the identity status of younger adolescents are reviewed. In the third section the two studies which have used the OhIEIS to assess identity status are described. STUDIES USING NEASURES OTIER THAN TIE MARCIA INTERVEW OR TIE OMEIS In one of the earliest studies of identity development, Howard (1960) administered a questionnaire of her own design to sixty-nine women who were sophomores or seniors at two high schools. She derived the questions from her readings of Erikson’s work, and included the areas of time distortion, identity consciousness, work paralysis, sexual identity, and diffusion of ideals. The area of identity consciousness area had five parts, including awareness of one’s identity, the extent to which one was building a self-image by copying other people, awareness of a sense of isolation, extent of one’s satisfaction with one’s present self and life, and the nature of one’s peer group relations. In another section of the questionnaire the nature of the women’s interpersonal l5 relations was addressed, as the purpose of the study was to investigate the relationship between identity and the psychosocial stage which follows it, intimacy vs. isolation. Her findings were that in most areas of conflict for which she tested, at least two questions were answered positively by ”twenty to thirty” girls, which she took to mean that there was a certain elevation of awareness of identity concerns among her sample. She also found that having conflict in one area was strongly associated with having conflicts in other areas, and she described such multiply conflicted individuals as ”identity diffused”. More girls were rated as being conflicted in the area of sexual identity than any other. Further, girls with some reported conflict with their parents were frequently more deeply engaged in identity issues. Finally, a trend toward lessening conflict in the older group (seniors) compared to the younger group (sophomores) was found, though it was not statistically significant. The weaknesses of the study were primarily the lack of any sort of validation for the instrument used, except for face validity, as well as the fact that response bias was not controlled for, and that the sample was limited in size and was only female. Nevertheless, there was some indication that there is a change in identity status during the high school years. Ciaccio (1971) used a Thematic Apperception Test-type story-telling instrument to assess the extent to which males aged 5, 8, and 11 expressed concerns associated with each of the first five stages described by Erikson. The results of a ”unit utterance” scoring procedure indicated that 11 year old boys were beginning to show some identity concerns, but the frequency of such expressed concerns was far lower than for concerns associated with the earlier stage components of industry and initiative. Although the study used only males, and the measurement instrument lacked many psychometric qualities, it has helped to suggest eleven or twelve years old as a possible lower age boundary for the study of normal identity development, and also supports Erikson’s view that some focus on earlier stages precedes consideration of identity issues. 16 Stark and Traxler (1974) used Dignan’s Ego Identity Scale to examine identity crystallization in late adolescence. Their sample consisted of college students, whom they grouped into two age groups, 17-20 and 21-24. Their purpose was to compare the two age groups, as well as to examine sex differences, and the effect of anxiety on identity development. Dignan developed the Ego Identity Scale from items adapted from self-report inventories and derived from Erikson’s concept of ego identity. The test consists of 50 items and has different forms for males and females, since it was originally designed for women and then expanded for use with men. In the original study (Dignan, 1963), it was found that female ego identity crystallizes in late adolescence, and that ego identity and manifest anxiety are significantly negatively correlated. Stark and Traxler found that 17 to 20 year olds reported significantly more ego diffusion than 21 to 24 year olds. Furthermore, they found that females in the 17 to 20 age range reported significantly less ego diffusion than males of the same age range, as did females in the 21 to 24 age compared to males in the 21 to 24 age range. Both males and females became increasingly more crystallized in their identity as they got older. Finally, they replicated the finding of Dignan that identity diffusion correlates significantly with anxiety. They interpreted these findings to mean that the significant crystallization of ego identity occurs in late adolescence, compared to early adolescence, that females resolve identity issues more rapidly than males, and that heightened anxiety prevents one, to some extent, from progressing to a crystallized identity. However, while showing that, as a group, 21 to 24 year olds are more identity crystallized than 17 to 20 year olds, they had no way of addressing to what extent the younger age group had begun to deal with identity issues, and moreover, did not attempt to say what percentage of each group might be ”high” or ”low” in crystallized identity. Thus, they left open the possibility that a significant number of younger subjects might have a crystallized (or crystallizing) identity, while the group as a whole still appears to be identity diffuse. 17 In a study of correlates of adolescent identity development, Lavoie (1976) administered a variety of measures to 120 sophomores, juniors, and seniors in high school. The extent of identity achievement was measured using the Marcia incomplete sentence blank test, called the Ego Identity Status Scale, which provides a score of overall identity achievement, though not one of identity status. Sex role identity was measured by an adjective check list called the Heilbrun Masculinity-Femininity Scale, in which high scores indicate greater masculinity and low scores greater femininity. Another measure of sex role identification was provided by a semantic diflerential measure developed by Lazowick in which a subject makes judgments of the similarity of the concept ”myself” to the concepts ”husband”, ”wife”, ”mother”, ”father”, ”man”, ”woman”, and ”family”. The Erikson Measure of Personality Development, developed by Constantinople (1969), was used as a measure of personality synthesis. This scale is a self rating scale which assesses personality integration on the first four psychosocial stages as well as the sixth (Intimacy vs. Isolation). It includes a measure of the fifth stage, identity, but this was omitted as the authors felt that this was already being assessed by the incomplete sentence blank. Self concept was assessed using the Tennessee Self-Concept Scale which also provided six measures of psychological adjustment (Defensive Positive, General Maladjustment, Psychosis, Personality Disorder, Neurosis, and Personality Integration). Finally, socialization practices were obtained from an 18 item child rearing questionnaire on which each subject rated each of their parents on the factors of warmth, concern, punishment practices, and consistency. Their results showed that while ego identity became greater with age, this increase was not statistically significant. Furthermore, there were no sex differences in ego identity. The other variables were analyzed by making a median split between high and low identity subjects and using that as one independent variable, with age and sex the other two. While high from low identity subjects could not be distinguished on a measure of masculinity / femininity, high identity subjects perceived less distance between self and father, self and family, self and man, 18 self and mother, and self and husband. The only significant sex by identity interaction occurred for self and family, where high identity females perceived greater similarity than low identity females. High identity subjects scored more positive on basic trust and industry than low identity adolescents. There were no significant sex differences on the personality measures. Subjects high in ego identity were better adjusted psychologically than low identity subjects, except in the area of Psychosis, in which there was no difference. There were no sex differences in psychological adjustment. Self concept was more positive with increasing age, and was significantly higher among high identity subjects than low scorers. Overall, the study supports the view that some increase in ego identity status may occur during the high school years, though it may be rather small. Another important finding was that successful resolution of the earlier stages of trust (Stage 1) and industry (Stage 4) predict more frequent resolution of identity (Stage 5) in early adolescence. An important non- finding was that in no case was a sex difference on a dependent measure associated with greater or lesser identity development, suggesting that sexual differences may be less important in early adolescence. Pomerantz (1979) studied the development of self-esteem, physical self-satisfaction, and identity, as predictors of the variable satisfaction with one’s social milieu. Her sample consisted of six hundred and nine junior and senior high school students in the eighth, tenth, and twelfth grades, evenly divided among males and females. Self-esteem was measured by the Rosenberg Self-Esteem Scale. Identity was measured by the Ego Identity Scale developed by Rasmussen (1964), which assesses the first six Eriksonian stages, all of which are presumed to be reworked by the adolescent in an identity crisis. Each stage is represented by three statements, and are rated on a four point Likert scale. The split-half and test-retest reliabilities for this measure are .85 and .87 respectively. The scale is reported to correlate significantly with Constantinople’s Inventory of Psychosocial 19 Development. Physical self-satisfaction was measured by the Body Cathexis Scale. Satisfaction with one’s social milieu was assessed using an Adolescent Questionnaire created by the authors, which asked adolescents to rate their degree of satisfaction with various aspects of their life. The best predictors of satisfaction with one’s social milieu were self-esteem, identity, and physical self-satisfaction in that order for males (accounting for 37% - 44% of the variance), while identity, physical self-satisfaction and self-esteem were the ordered predictors for females (42% - 47% of the variance). In each case, the first variable entered into the prediction equation accounted for the overwhelming bulk of the variance, while the other two variables accounted for 4% - 7% of the variance together. Thus, identity was found to vary significantly with age for women, but not for men, in early adolescence. That result was tempered by the fact that the variables identity and self-esteem correlated more strongly with each other than either did with the criterion variable. Thus, the possibility exists that the effect of one variable is masking the effect of the other, but the authors of the study do not perform the analyses necessary to determine if that is the case. Erwin (1983) investigated whether there was a difference in identity level between students planning to go to college, those planning not to go to college, and those who were undecided. His test of identity was based on Erikson’s ideas, and was divided into three areas: Confidence, Sexual Identity, and Conceptions about Body and Appearance. The author developed a questionnaire to assess these constructs, which were shown to have internal consistency coeflicients ranging from .75 to .81. The scale is also reported to correlate moderately with other scales of personality integration and internal locus of control. The scale was administered to 163 high school students from grades 9 through 12. Those students who were sure they either did or did not want to go to college and those who were uncertain about their'plans could be difierentiated by both the subscale ”Confidence” and the subscale ”Conceptions about Body and Appearance”. The students who were sure about 20 their choice appeared essentially identical on all measures, consistently higher than those who were unsure. No study was made of sex or cohort differences. STUDES USING TIE MARCIA INTERVEW TO ASSESS IDENTITY STATUS The Marcia semi-structured interview, by far the most frequently used measure to assess identity status, was first used on a sample of students younger than college age in 1978, when Raphael (1978) tested one hundred women in the 12th grade. In addition to identity status, he also administered measures of social class, intelligence, tolerance of ambiguity, cognitive complexity, ego identity, and anxiety. He also examined performance on a behavioral measure of information search. Significantly, the author dismissed on theoretical grounds the possibility of a high school senior being in the identity achieved status. To be in that status, one must have considered alternatives and committed one’s self to one of them. At that age, the author felt that a subject would of necessity have considered too few alternatives, and thus be actually identity foreclosed, or not have made a true commitment, and thus be actually in the moratorium status. Therefore, every person in this study was defined as either identity diffused, foreclosed, moratorium or unclassifiable. By these rules, 22 persons were found to be in moratorium, 32 in foreclosure, 24 in the identity diffusion status, and 22 were unclassifiable. Moratoriums were significantly higher than foreclosures on the variables social class, complexity of conceptual level, intelligence, and tolerance of ambiguity. Moratoriums were significantly higher than diffusions on the same variables, as well as certain measures on the information search task. Foreclosures were significantly higher than diffusions only on the measure of ego-identity. The Rasmussen Ego identity scale was used, giving only an overall score on ego identity. Foreclosures were higher on the ego identity variable than moratoriums, though not significantly so. 21 The author concludes that the results support the usefulness of the identity status categories, as well as the hypothesis that the moratorium status is a more adaptive one in terms of overall behavioral functioning than the foreclosure status, while the foreclosure status, although more rigid and less tolerant of change, is still more behaviorally adaptive than the identity diffusion status. A significant percentage of subjects in a high identity status was observed, although only women were tested, and no identity achievement status was scored for. In perhaps the only study which examined the identity status of young adolescents and of older adolescents, Meilman (1979) administered a modified Marcia Identity Status Interview to twenty five white males in each of five age groups: 12, 15, 18, 21, and 24 years. A series of questions designed to elicit subject’s exploration of and commitment to aspects of sexual identity was added and administered to the three oldest age groups. A series of questions related to avocations was also asked, but were poorly correlated with overall crisis and commitment scores, so were excluded from further study. A multivariate analysis of variance of the effects of age on commitment and crisis measures for the areas of occupation, religion, and politics was significant. Univariate tests indicated age-related increases in occupational commitment and crisis, religious crisis, and political commitment and crisis. The effect of age on sexual commitment and crisis was also significant. Cross sectional data showed that there were steady decreases with age of subject’s presence in the ”lower” identity statuses (identity diffusion and foreclosure), and a steady increase with age of a subject’s presence in the ”higher” identity statuses (moratorium and particularly identity achievement). Of interest is that the author classified 4% of 15 year olds as identity achieved, and found fully 20% of 18 year olds in that category. There was no evidence that, as a group, higher identity statuses regress to lower ones, after college, but no analysis of such changes within individuals could be made, and that is the relevant variable. Although this was a cross sectional study, it suggests that while the majority of identity resolution is done 22 during the college years for men, a significant minority become identity achieved by the end of their high school years. St. Clair and Day (1979) tested the hypothesis that adolescent women who are high in ego identity have high interests in religious and political values. In addition to the Marcia Interview, they administered the Allport-Vernon-Lindzey study of values, and collected data on the parents’ occupational, educational, and marital status, as well as obtaining the subjects’ GPA (grade point average) and IQ scores on the Otis-Lennon Mental Ability Test from school records. The age of the women was either 17 or 18 years. Of the 80 subjects, 40 were classified as being in the moratorium status, with 15 identity achievements, 14 foreclosures, and 11 diffusions. Of the values investigated, only religious values were difl'erent across statuses, with diffusions having significantly lower interest in religious values than any of the other three statuses. Furthermore, diflusions had significantly lower GPA’s than achievements, moratoriums, or foreclosures, but the IQ’s of the four identity statuses were not significantly different. An interesting finding was that 67% of the identity achieved women came from families in which either divorce or death of one parent had occurred. There was also a significant relationship with mothers’ educational level, in which more than half of the mothers of achievements and foreclosures had attended college, 30% of mothers of moratoriums had done so, and only 9% of mothers of diffusions had. There was no relationship with level of father’s education. Thus, the authors found that 19% of their high school senior subjects could be assigned to the identity achieved status. The difference in GPA between achievements and diflusions without a corresponding difference in IQ’s suggested that one difference between the two groups was in the area of motivational or interest variables rather than aptitude. There was also an indication that instability in the family may facilitate the identity achievement of women, forcing them to consider alternatives and make commitments sooner than others. 23 In a comprehensive study of the identity development of younger adolescents, Archer (1982) assessed the identity status of 6th, 8th, 10th, and 12th graders. She was careful to include equal numbers of men and women, and obtained information about their parent’s educational and vocational level. Rather than using the interview to assign an individual to one identity status, the areas of vocational, religious, political, and sex-role interests were each scored separately so as to examine each one’s importance. The majority of subjects used diffusion and foreclosure ”decision-making strategies”, but a significant increase of those in the achiever status as grade increased was found. No significant sex differences were found throughout the study, although the greater number of males in the foreclosure status approached significance. Among the four different content areas of investigation by the interview, there were far more diffusions in the political philosophy area than any other, with almost none in the sex-roles area. By contrast, the sex-roles area was primarily characterized by a very high number of foreclosures. There were few individuals in the moratorium status, but they were largely concentrated in the vocational plans area. The identity achievers were equally found in the vocational and religious areas, with many fewer in the sex- role or political areas. Presence in the achiever status correlated significantly with both parents’ educational and vocational levels, i.e., the more highly educated parents with the best jobs tended to have sons and daughters who had considered alternatives and made commitments to one of them. Perhaps the most interesting result to emerge from the study was the sense that individuals are in different identity statuses for different areas of their lives, and are usually in more than one decision-making mode at once. Thus, only 5% of the subjects were in only one status, 60% were in two, 32.5% were in three, and 2.5% were in all four statuses. A strong question was thus raised about the advisability of labeling an individual with only one identity status. 24 In summary, Archer found a significant shift toward use of the achiever status in early adolescence, almost exclusively in the areas of vocational plans and religious values. Questions regarding politics were generally not under consideration, while commitments without. awareness of choices were made regarding sex roles. No sex differences were found throughout the study. Parents with more education and higher level occupations tended to have identity achieved children, and an individual was in one identity status across all four content areas only 5% of the time. In a recent study investigating the identity development of early adolescents, Gortych (1983) measured the identity status of 27 ninth grade and 25 twelfth grade women. She also measured their ego strength, using the ego-strength scale of the MIVIPI. Her hypotheses were that there would be a shift from the diffusion status to the moratorium status as women got older, and that this shift would be accompanied by an increase in ego strength. The majority of her subjects, whether 9th or 12th grade, were in the moratorium status, with none in the achiever status, few in foreclosure, and 10% - 20% in diffusion. There was, however, a significant difference in ego strength among the 12th graders (higher ego strength) compared to the 9th graders. Thus, in her sample, Gortych found essentially no change in identity status between 9th and 12th graders, with no one in the achiever status. STUDES USING TIE OBJECTIVE MEASURE OF EGO IDENTITY STATUS (ONEIS) There have only been two studies in which the identity development of young adolescents was investigated using the ONEIS as the measure of identity status. In one, Adams and Jones (1983) assessed high school women from the 10th, 11th, and 12th grades. In addition to recording their identity status, they also examined the perceived parenting styles which the adolescents felt they were living in. The questions tapping parenting styles related to issues such as perceived fairness of punishment received, extent of parental control, extent of allowed 25 freedom, amount of expressed approval and praise, and frequency of expressed love and affection of parents toward their children. The authors found evidence for the validity of the OhIEIS when used with a younger sample. Thus, correlation coefficients showed the diffusion and achievement statuses to be strongly negatively correlated, as expected, as well as moderate correlations between diffusion and foreclosure, and difl'usion and moratorium, which seem to indicate that the issue of commitment is more strongly tapped than that of crisis, or the consideration of alternatives. No significant differences were found in the examination of identity status change by grade or age. The authors report two ”meaningful but nonsignificant trends” among the four age groups (15 to 18 years, inclusive), in which 15 year olds and 18 year olds had higher foreclosure scores, while 16 year olds and 17 year olds had higher achievement scores. With respect to the parenting variables, adolescent girls in the ”higher” statuses of moratorium or achievement both involving an active state of thinking about alternatives, were the least likely to report controlling behavior on their mother’s part, instead seeing their mothers as encouraging independent behavior. Further, fathers were viewed as fair in punishment, while offering minimal approval and praise to their daughters. The authors conclude that their data are somewhat equivocal in terms of a developmental trend of identity in middle adolescence. The expected relationship was ruined by the oldest age group showing lower achievement and higher foreclosure scores than predicted by the trend. The difference is attributed either to a cohort difference or to the special pressures attending an impending high school graduate. Overall, however, diffusion scores decreased with age, and achievement scores increased. In a study relating identity status to locus of control, Abraham (1983) tested the hypothesis that subjects in the achievement status feel that reinforcements are internally controlled. The subjects were 254 boys and girls with roughly one fourth each in the 9th, 10th, 26 11th, and 12th grades. In addition, because 30% of the sample was Mexican-American, the opportunity was taken to investigate ethnic heritage as a variable. Finally, one of the scoring rules of the OMEIS was examined, in which subjects who score below the cutoff score for all four statuses, and cannot be assigned to a stage, are ”arbitrarily” assigned to the moratorium status, under the presumption that they are experiencing a peculiar form of crisis (See Chapter 3). These subjects are assigned to the ”moratorium rule 2” status. Twenty nine subjects were in diffusion, 29 in foreclosure, 19 in moratorium, 33 were identity achieved, and 122 were below all cutofl' scores and were in the ”moratorium rule 2” status. Identity achieved individuals were found to be significantly more internal in their locus of control than individuals in any other status. Also, ”moratorium rule 2” individuals were more internal than regular moratoriums. The ethnic differences were significant, and reflect the fact that Anglo-Americans were more numerous in the foreclosure and ”moratorium rule 2” status. The author concludes that internal locus of control is characteristic of identity achievers in early adolescence. Further, the ”moratorium rule 2” individuals must be considered to be in a qualitatively different state than other moratoriums, since they are two distinct groups on at least one intrapersonal variable. The difference is particularly important since ”moratorium rule 2” is a highly represented status, 55% in this study. The finding that Mexican-Americans had less of a tendency to be in the diffusion status went against the author’s assumption that parents’ socio-economic status would be a mediating variable. A post hoc analysis showed that the lower educational status of the parents of Mexican-Americans did not predict high presence in the diffusion status. The author concludes that parenting style may be a more important mediating variable. 27 SUMMARY The main findings regarding the development of identity status in early adolescence are summarized in Figure 3. Although the bulk of early adolescents are generally found in the ”lower” identity statuses, characterized by not considering alternatives, a fairly constant percentage, usually 10% - 20%, are generally found in the ”higher” identity statuses, characterized by consideration of alternatives and commitment to one of them. Usually the commitment is in the area of vocational choice and religious issues, more so than politics. On numerous measures of psychological functioning and adaptation, those in the higher statuses appear to use more complex reasoning, to be more adaptable, and more resistant to stress. The causal relationships among these variables are far from sorted out, however. Those studies that have examined sex differences in identity development, have generally not found them. With one exception, all of the studies were cross sectional designs, and examined white suburban or urban adolescents. Therefore, cohort effects may be being observed, and limits generalizability to the white middle class. That caveat is significant for there is evidence both that different ethnic groups develop their identity in different ways, and that the educational and occupational level of the parents is a significant variable in the identity development of their children. The most significant mediating variable was parenting style, with a more controlling, punitive style predictive of lower identity statuses. Parental conflict could actually help adolescents develop more quickly, perhaps by making them commit to one of several alternatives sooner than children of less conflicted households. Finally, there was some support for Erikson’s view that resolution of the earlier stages of trust vs. mistrust, and industry vs. inferiority in particular, are important to a successful resolution of the identity crisis. The findings have some drawbacks, however. In the first place, sample sizes are generally fairly small, a handicap which limits internal validity and, particularly, 28 Researcher Howard, 1960 Ciaccio, 1971 Stark & Traxler, 1984 Lavoie, 1976 Pomerantz, 1979 Erwin, 1983 Identity Instrument Her own ques- tionnaire Story telling test Dignan Ego Identity Scale Ego Identity Status Scale Rasmussen Ego Identity Scale Erwin Identity Scale Sex of Ages Sample If 14-17 M 5,8,11 M,F 17-24 M,F 15-18 M,F 16-18 ih4,F‘ .14-153 Main Findings Evidence for some change with age, pri- marily in sexual identity. 11 year old boys have early identity con- cerns. Identity crystal- lizes at 21-24 years. Women resolve identity issues more quickly than men. Nonsignificant increase in iden- tity with age. Resolution of earlier stages predicts resolu- tion of identity. No sex differences. Identity varied with age for women, not for men. Students sure about plans for college had higher identity than those un- sure. Figure 3. Summary of Research on the Identity Status of Early Adolescents 29 Researcher Raphael, 1979 Meilman, 1979 St.Clair & Day, 1979 Archer, 1982 Gortych, 1983 Identity Sex of Instrument Sample Marcia Identity F Status Inter- view (MISI) Rasmussen Ego Identity Scale MISI M MISI F MISI M,F MISI F Figure 3. (continued) 30 Ages 17-18 12,15,18,21,24 17-18 11-18 14-18 Main Findings Moratoriums more adaptive than Fore clo- sures, and both more adaptive than Diflusions Significant number in high identity status. Identity status increased with age. 20% of 18 year olds identi- ty achieved. 19% of subjects identity achieved. Fam- ily instability may facilitate identity resolu- tion. Increase in number of sub- jects in achieve- ment status with age. No sex diflerences. No individual in achievement status. Researcher Adams & Jones, 1983 Abraham, 1983 Identity Sex of Instrument Sample Objective Meas- F ure of Ego Iden- tity Status (OMEIS) OMEIS M,F Figure 3. (concluded) Ages 15-18 14-18 Main Findings No identity status change with age. Non controlling parenting style facilitates higher identity status. 15% of subjects in achievement status, with internal locus of control. generalizability. Further, the bulk of recent work in identity development has been done using the Marcia Identity Status Interview. The Identity Status Interview measure is time consuming and labor intensive. Moreover, it has been criticized by its designer for possible loss of objectivity (Marcia, 1976). The development of the Objective Measure of Ego Identity Status, designed to help redress those drawbacks, has not done so, because only two studies using a younger adolescent population have been conducted. Those studies suflered from test development problems which may have clouded the results, which were not consistent with the bulk of previous research findings. Therefore, the need for a valid, objective measure of identity status, which could examine the identity development of younger adolescents remains as a prime measurement need in developmental research. 31 CHAPTER 3 NETHODOLOGY In Chapter Three, the process of obtaining a sample is described, as well as the instruments used in the study, the design of the study, and the method of analysis. The hypotheses are restated in testable form. SAMPLE AND POPULATION The population under consideration was younger adolescents, specifically defined as individuals not yet having entered college. The desired age range was 13 to 18 years. A number of junior high school (middle school) and high school principals in the Greater Lansing area were contacted, with the principals of two of them agreeing to have their schools participate in the study. The sample was composed of those students who had signed permission forms from their parents, and who were present on the day the instruments were administered. Both communities may be characterized as moderate sized towns, middle to upper- middle class, and quite predominantly white. Both school’s principals indicated a low drop-out rate of about 2% - 3%, and indicated that at least 95% of a given class eventually graduates. The sample had the following demographics: the total sample was 214, with 52 eighth graders, 98 tenth graders, and 64 twelfth graders. Males composed 49.5% of the overall sample, while females composed 50.5%. There were 24 13 year olds (11%), 29 14 year olds (14%), 87 15 year olds (41%), 11 16 year olds (5%), and the age of the last 29% of the sample was unknown. The sample was overwhelmingly white, as 95% of those subjects for whom race was known were white, with only three members each from the black and hispanic races. INSTRUMENTS Besides the ”Fact Sheet”, used to collect demographics on the sample, two clinical instruments were administered: the Objective Measure of Ego Identity Status (OIVEIS) and a shortened version of the Assessment of Adult Adjustment Patterns (AAAP). Fact Sheet The fact sheet collected the following data from all students: grade in school, and gender. Because a different version was used for each of the three grades, the students’ vocational intentions after high school (and marital intentions after high school) were collected only for the eighth graders, and race and age were not collected for the twelfth graders. Examples of the fact sheets may be seen in Appendices B and C. Objective [Measure of Ego Identity Status (OAIEIS) The OIVEIS was developed by Adams, Shea, and Fitch (1979) as a paper and pencil test to establish which identity status an individual currently occupies. The test was developed out of a need to have an objective, yet valid test for ego-identity status, designed to replace the cumbersome interview schedule and incomplete sentence blank already in use. The interview schedule had been criticized by its designer for a lack of consistency arising primarily because of difficulty in scoring portions of it (Marcia, 1976). The OMEIS was also designed to allow a researcher to measure the extent to which a given subject tends toward self-perceived identity difiusion, foreclosure, moratorium, and achievement separately, instead of simply assigning a subject unilaterally to one status, as do the interview and the incomplete sentence blank. The ONEIS is structured with 24 items. For each item, one must indicate whether the statement is similar to or dissimilar to one’s self-perceptions, by marking a number on a Likert scale with six gradations ranging from ”strongly agree' to ”strongly disagree”. Each identity status has six questions designed to assess the extent to which an individual responds 33 like others in that status, with two questions each drawn from the psychosocial realms of occupation, religion, and politics. The test is scored by assigning a value of ”l” to an answer of ”strongly disagree” to any question, a value of ”6” to an answer of ”strongly agree” to any question, and the appropriate intermediate values to the other answers. The values of the six questions pertinent to each identity status are then summed, giving four scores, or an ”identity status profile” for each subject. Operationally, the subject’s identity status may be determined by calculating the means and standard deviations of each identity status for the entire sample under consideration, then assigning each individual to an identity status according to the following directions: 1. Individuals with scores falling one standard deviation above the mean for a given identity status are scored as being in that identity status if all remaining scores fall below the one standard deviation cutofl'. 2. Individuals with all scores falling less than one standard deviation above the mean are scored as moratorium. The rule was arbitrarily decided upon, as a low profile of this sort was assumed to reflect a unique form of crisis. Subsequent research has shown, however, that individuals in ”moratorium rule 2” are going through a qualitatively different experience than those scored as )3 being in moratorium by rule ”1 above. 3. Individuals with more than one score above the standard deviation cutoff are scored as being in transition from one status to another and are given a transition status typology, such as diffusion-moratorium, difl'usion-foreclosure, etc. The normative group consisted of 48 participants (12 male, 36 female) in a general education course at Utah State University. They were primarily middle class youth from rural or 34 small town backgrounds. Their academic fields were education, child development, sociology, social work, psychology, and liberal arts. The OMEIS was partially validated by its developers through four studies comparing its results with those obtained from the Incomplete Sentence Blank (ISB) developed by Marcia (1966), which itself provided the primary validation measure for the ego-identity interview . The OMEIS had internal consistency coefficients for each status ranging from .67 to .76. A one way analysis of variance using the ISB scores as a dependent measure yielded a significant result. Individual comparisons using a multiple range test revealed that diffusion status persons were significantly less committed to an overall ego identity as measured by the ISB than the identity achievement status persons. A second study supported the ability of the OIV‘IEIS to categorize individuals in the same manner as the ISB, as a second one way analysis of variance with a slightly different sample again proved significant. Concurrent validation with a measure of authoritarianism supported the expectation that foreclosure individuals are significantly more authoritarian than moratorium or achievement persons. Concurrent validation with a measure of self acceptance supported the expectation that achievement individuals are significantly more self accepting than individuals in diffusion or foreclosure. Concurrent validation with a measure of personal rigidity supported the expectation that foreclosure individuals are more inflexible (though not significantly so) than achievement individuals. A third study again supported the OMEIS as able to categorize individuals similarly to the ISB. Furthermore, gender effects were investigated, and were consistently found to be non-significant, that is, women and men score the same way on this test. A cross sectional study by age showed results in the expected direction: younger individuals were more diffused and foreclosed, while older ones were more likely in moratorium or achievement. 35 The fourth study compared results of the OMEIS with the Marcia semi-structured interview. Although the results were generally as expected, there was some discrepancy in the classification of an individual into either the diffusion or moratorium classes, as several subjects were transposed between the two classes. Two explanations are offered, one, that the OMEIS may not be completely adequate as a tool to distinguish between identity diffusion and moratorium, or two, that such individuals were in a ”diffusion-moratorium transition state” and thus could be classified as either status depending on the specific criteria examined. Test-retest. reliability ranged from .71 to .93 for the four subscales. Taken together, these four studies appear to offer evidence for a certain amount of concurrent validity for the OMEIS, particularly with respect to the theoretical formulations and research of James Marcia. Assessment of Adult Adjustment Patterns (AAAP) The AAAP was recently developed by Farquhar, Wilson, and Parmeter (1977) as an attempt to objectively measure an individual’s degree of mastery of each of the eight psychosocial tasks postulated by Erikson. In its most complete form, the AAAP consists of 320 items, including items for each of the eight stages, as well as for two validity scales. Each item consists of a statement which a respondent describes as ”Definitely true of me”, ”True of me”, ”Not true of me”, ”Definitely not true of me”, or ”Does not apply to me”, by marking a number on a 5-point Likert scale. Items may be presented either ”positively”, meaning that an endorsement of the item as ”True of me” is supportive of having mastered the stage, or ”negatively”, meaning that a non-endorsement of the item as ”Not true of me” is supportive of having mastered a given stage. The test is scored by giving a respondent four points for an item if they have answered in the manner most supportive of their having mastered the stage, i.e., ”Definitely true of me” for a ”positive” item, or ”Definitely not true of me” for a ”negative” item. The item is given three points if a respondent has endorsed the next most 36 supportive statement, and two points or one point for the least supporting statements. The item is scored with no points if a respondent has endorsed it as ”Does not apply to me”. All item scores pertaining to a given stage are then summed to give a stage score. James Azar, in an analysis of the AAAP using a norming population of university students and faculty, established cutoff scores for each stage (Azar, 1983). If the score on a given stage is above the cutoff point, it indicates that an individual has answered a substantial portion of the items in the mastery direction, and is presumed to have resolved to a significant extent the psychosocial crisis of that stage. The cutoff points for the stages were set so that they had the characteristics of Guttman scales, specifically, that they were mastered in hierarchical order of difficulty, beginning with Stage 1 (easiest) and proceeding in order to Stage 8, the hardest scale on which to achieve a mastery score, and thus the last one to be mastered. The scales of the AAAP have internal consistencies ranging from .88 to .95 and were normed on university students and faculty. It gives results in the expected direction for a variety of populations, including psychiatric inpatients and a prison population. However, there have not been any direct studies of the validity (concurrent or predictive) of any of the AAAP scales. For the current study, only the first five scales of the AAAP were used to collect data regarding adolescents, partly because the questions relating to intimacy, generativity, and old age were assumed not to be relevant to adolescents who are predominantly wrestling with much earlier issues. Another concern was that the length of the full AAAP, 320 items, would make the time necessary to administer all of the instruments equal to at least two class periods, a sacrifice of time no principal was willing to agree to for his students. One of the two validity scales, an F-scale patterned on the NIMPI scale of the same name, was also dropped, but the other, a social desirability scale, was retained, and brought the total number of items on the shortened AAAP to 205. 37 The social desirability scale was used to identify subjects who had a strong tendency to respond to items as they felt they were expected to, or how they felt others might respond, rather than according to their own feelings. Recognizing that different levels of social desirability existed for each of the three cohorts, an average social desirability score and standard deviation was computed for each of the three grades sampled. Those students were eliminated from subsequent analyses who had social desirability scores greater than one standard deviation above the mean for their grade. The procedure resulted in the elimination of 29 cases from the overall sample. More detailed discussion of the effect of the exclusion procedure may be found in Chapter 4. PILOT STUDY A small pilot study was conducted to see how understandable the test materials were, how long it took to complete them, and to check the scoring procedures. Four adolescents aged 15 to 17 years were recruited through an acquaintance network. They were, as a group, given an explanation of the nature of the research and what they were being asked to do. They were given a packet of materials containing the fact sheet, the Objective Measure of Ego Identity Status, the modified and shortened version of the AAAP, and a computer scoring sheet for recording their AAAP answers. There were no questions regarding the instructions or any part of the test materials, except for one comment that questions on the AAAP appeared to be redundant and in some cases identical. This comment reflects the nature of the test, which was designed with a great deal of redundancy. In the final version of the test, the redundant versions of two questions, which were found to be exact duplicates, were dropped. The time required for completion of the test materials ranged from 31 minutes to 44 minutes and averaged 37 minutes. Thus, the materials were deemed suitable for use in schools where a class period is from 45 to 50 minutes long. The OMEIS was tabulated and scored by hand, and revealed one student in the moratorium status of identity, one in foreclosure, and one 38 in difiusion. The final student was below the cutoff points on all statuses, and was thus assigned to the ”moratorium Rule 2” group (see the ”Instruments” section of this chapter). The AAAP, because of numerous items and complex scoring procedures, must be scored by computer. A standing program exists in the Scoring Office of Michigan State University’s Computer Center which can accomplish this. However, difficulties arose in its use, which were not resolved by the time the data from the schools was due to be collected, so pilot study AAAP results were not obtained. ADMINISTRATION OF THE MATERIALS After approval was given by the principal for the project to proceed, parental consent forms were sent to each school to be distributed to the students and taken home. Approximately one week later, when most of the forms had been returned, materials were dropped off at the office of the principals for distribution to each involved teacher. In both schools, the principals felt that more effective data collection could be accomplished if the teacher normally scheduled to teach a class presented the materials. Usually, these were Social Studies or English teachers. After the materials were returned to the principal’s office of each school, they were held there until picked up by an associate. DESIGN OF THE STUDY The study was cross sectional and exploratory in nature. The independent variable was grade in school, with three levels: 8th grade, 10th grade, and 12th grade. One dependent variable consisted of the identity status scores obtained from the Objective Measure of Ego identity Status, which had four levels: identity diffusion, foreclosure, moratorium, and identity achievement, and the other dependent variable consisted of the stage scores obtained from the AAAP, and had five levels, one corresponding to each of the first five Eriksonian stages from 39 trust through identity. For one analysis, the scores on the items of the four identified factors of the fifth stage, identity versus role confusion, were used as separate dependent variables. HYPOTHESES IN TESTABLE FORM In the following section, each of the research hypotheses are restated in turn, in words and symbolic form, and include null and alternate hypotheses. The hypotheses relating to the validation of the OMEIS with a younger sample were: Hypothesis 1 H H 01‘ Al: The proportion of students at each grade level in identity diffusion remains the same during early adolescence. The proportion of students at each grade level in identity diffusion decreases during early adolescence. Stated symbolically the hypotheses are: H H Hypothesis 2 H 01: A1: A2: D8=D10=D12 where : D8 = the proportion of 8th graders in identity diffusion, D = the proportion of 10th graders in identity diffusion, 10 D12 = the proportion of 12th graders in identity diffusion The proportion of students at each grade level in moratorium remains the same during early adolescence. The proportion of students at each grade level in moratorium increases with grade level during early adolescence. Stated symbolically the hypotheses are: H 02‘ MS=M10=M12 4O H M8 D10 > D12) to provide an increase in power, in the event that the presumed ordering actually exists. If the proportions are not ordered, then the technique, which requires one to average sets of proportions which do not conform, increases the likelihood that no significant effect will be found. The second two hypotheses of the first set were tested by a more general procedure which tests only for differences among subgroup proportions compared to the overall sample proportion. Analysis ofhypothescs 5 through 8. The second set of hypotheses, with the exception of Hypothesis 8, were analyzed using a fourfold table of proportions. In such a table, one assesses the presence or absence of association between two characteristics by observing the frequency with which they are each found or not found for a sample of individuals. Each individual is assigned to one of four cells, depending upon whether they have both characteristics of interest, have the first but not the second, the second but not the first, or neither. The test statistic is a chi-square with one degree of freedom, and is generated by comparing the observed cell frequencies (or proportions) with the frequencies expected under conditions of total independence of the two characteristics. The eighth hypothesis was tested by canonical correlation. This is a technique which gives one measure of the amount of intercorrelation between two sets of variables. The variables within each set are presumed themselves to be intercorrelated, and each set of variables is presumed to correlate with the other set. The process of canonical correlation then produces linear combinations of each set of variables (called canonical variates), such that the correlation between the canonical variates is maximized. The square of the canonical correlation, the eigenvalue, represents the amount of variance of one set of variables accounted for by the other set. In addition, coefficients for each variable in the canonical variate may be obtained, and examination of these can lead to the formation of hypotheses relating to the relationship between particular variables in the canonical variates. 44 Post-hoe analyses. Because of the nature of the data and the purposes of the study, several additional analyses were performed: a Guttman scale analysis of the AAAP as it was used in this study, the determination of new mastery cutoff points for the AAAP, the reanalysis of hypotheses 5, 6, and 7 using the new cutoff points, and a MANOVA (Multivariate Analysis of Variance) on the OthIS data. Guttman scale analysis is a set of methods for evaluating whether a set of scales have the properties of Guttman scales. Those properties are first that the scales must be unidimensional, that is, measure movement toward or away from a single underlying concept. Second, the scales must be cumulative, that is, the scales are ordered from ”easiest to pass” to ”hardest to pass”, and moreover, a respondent who endorses positively a difficult item should endorse positively all easier items. In the present study, the five scales analyzed were those scales pertaining to the five stages of development in the AAAP. The stages were analyzed twice, once allowing them to be ordered from easiest to hardest, once forcing them to be ordered the way they are conceptualized (i.e., Stage 1 first, then Stage 2, etc.). In each analysis, several statistics were computed, including the coefficient of reproducibility, a measure of the extent to which a respondent’s score is a predictor of one’s response pattern, and the minimum marginal reproducibility, which is the minimum coefficient of reproducibility that could have occurred for the scale. The percent improvement is the difference between the first two measures and indicates the extent to which the coefficient of reproducibility is due to response patterns rather than the inherent cumulative interrelation of the variables used. The coefficient of scalability is a measure of the extent to which a scale is truly unidimensional and cumulative. Guttman scale analysis can also be used to examine how the qualities of a set of scales change as the cutoff points, or the levels at which each scale in a set is considered to be passed, change. In the case of the AAAP, the cutoff points were the mastery levels determined by Azar (1983). Since use of those mastery levels with the current sample led to many subjects mastering none of the stages, new cutoff points were determined for the AAAP so that a much 45 greater number of subjects had mastered at least one of the scales. The criteria for the new cutoff points were that they had to be such that they allowed most of the subjects to be assigned to one of the first five stages of the AAAP, while keeping the coefficient of reproducibility and the coefficient of scalability high. Further, the number of ”errors” was kept as low as possible. An ”error” is a case in which a subject passes a high scale while failing a lower one, i.e. deviates from a pure Guttman pattern. The next post hoc analysis involved retesting hypotheses 5 through 7 using the distribution of subjects obtained with the new mastery cutoff levels. The same cutoff points for the OMEIS were used. Multivariate Analysis of Variance (MANOVA) is a technique which allows a researcher to test the significance of differences between several interrelated dependent variables measured at an interval level, across one independent variable, measured at a nominal level. In the present study, the independent variable was grade level, while the four dependent variables were transformed scores of the four subscales of the OMEIS. The scores were transformed to investigate changes in the presumed underlying variables tapped by the OMEIS. Specifically, the four subscales of the OIVIEIS (identity diffusion, foreclosure, moratorium, and identity achievement) are theoretically derived by assessing the presence or absence of a sense of crisis (consideration of alternatives) and commitment in three psychosocial arenas (see Chapter 1 - ”Marcia’s Four Identity Statuses”). The subscale variables were transformed so as to obtain the following variables: a grand mean of all scores, two ”crisis” variables, formed by averaging the two subscales involving the presence of crisis (moratorium and identity achievement) and the two which do not (diffusion and foreclosure), two ”commitment” variables, formed by averaging the two subscales involving commitment (foreclosure and identity achievement) and the two which do not (diffusion and moratorium), and an interaction term assessing the extent of interaction between the two 46 variables ”crisis” and ”commitment”. The MANOVA was performed using as dependent measures the grand mean, a contrast formed by the difference between the high and low crisis variables, a contrast formed by the difference between the high and low commitment variables, and the interaction term. The final variables were constructed so as to be orthogonal. Both an overall multivariate significance test and univariate tests of significance were done with the transformed variables. SUlNIMARY In Chapter 3 the nature of the instruments used in this study, their psychometric qualities and validation work, and the techniques used to score them were presented. It was concluded that while the OMEIS has had some adequate validation work, the AAAP, while a well designed psychometric instrument, lacked concurrent and predictive validation work. The methods of administration of the test materials and the results of a small pilot study were reviewed. Two sets of test hypotheses were presented, including those pertaining to the validation of the OMEIS with a younger sample, and those pertaining to the validation of the AAAP using concurrent results from the OMEIS. The methods of analysis were reviewed, and several additional analyses were described, including the determination of new mastery level cutoff scores for the AAAP with the current sample, and the reanalysis of three of the hypotheses using the resulting new subject distributions. In Chapter 4 the results of the analyses are presented. 47 CHAPTER 4 RESULTS In this chapter, the results of the study are examined, beginning with the description of students by each instrument, followed by the outcome of testing each hypothesis. DISTRIBUTION OF CASES ON THE OBJECTIVE thASURE OF EGO IDENTITY STATUS The descriptive statistics for the OMEIS were calculated for each identity status. and may be seen in Figures 4 and 5. Mean Standard Kurtosis Skewness Deviation Identity Diffusion 21.593 4.538 .849 .122 Foreclosure 17.107 5.052 -.554 -.241 Moratorium 20.290 4.341 .847 -.185 Identity Achievement 23.864 4.155 .455 .252 Figure 4. Descriptive statistics over the whole sample (n=2l4) for the OMEIS These statistics were used to assign subjects to a given identity status, as described in Chapter 3, and resulted in the distribution seen in Figure 6. Of the entire sample (n=2l4), 80 (37%) could be unambiguously assigned to a single identity status. An additional 24 (11%) were in a transition status (e.g., diffusion-foreclosure). A further 7 (3%) could not be assigned to any category because of high scores on 3 or more identity statuses. Finally, 103 (49%) could not be assigned to any category because of low scores on all identity status scales. 48 Identity Diffusion Identity Diffusion Foreclosure Moratoriu m Identity Achievement 1.0 .1039 p= .065 .4731 p=.001 -.1553 p=.012 Foreclosure Moratorium Identity Achievement 1.0 .2719 1.0 p=.001 -.2236 -.3469 1.0 p=.001 p=.001 Figure 5. Pearson correlation coefficients of the identity statuses of the OMEIS over the whole sample (n=2l4) Identity Status D F M A 8 3 ll 4 6 Grade 10 7 8 4 9 12 3 6 4 15 D = identity diffusion F = foreclosure M = moratorium A = identity achievement n=80 Figure 6. Number of students in each identity status by grade level as classified by scores on the OMEIS DISTRIBUTION OF CASES ON THE ASSESSMENT OF ADULT ADJUSTMENT PATTERNS The distribution of students on the AAAP may be seen in Figure 7. 49 Stage Grade 10 30 1 6 2 2 12 15 4 l 1 l n = 84 Figure 7. Number of students in each Eriksonian stage by grade level, classified by mastery level at each stage Of the entire sample (n=209), 84 (40%) could be unambiguously assigned to a given stage. An additional 50 (24%) had ambiguous scores, in which one or more of the stage scores were below the cutoff point (e.g., scoring above the cutoff on stages one and three, but below the cutoff on stage two). Finally, 75 (36%) had the ambiguous score in which all stage scores were below the cutoff points. The reliabilities of the OlVfEIS and the AAAP were calculated for each of the three grades and are listed in Figure 8. EffECT OF THE CORRECTION FOR SOCLAL DESIRABILITY When the subjects who responded in the most socially desirable way were eliminated from the data as described in Chapter 3, the results of the OMEIS were modified as shown in Figure 9. Figure 9 reflects the fact that 7 cases were eliminated from the eighth grade sample, including one diffusion, two foreclosures, one moratorium, and three whose scores were low on all scales. 13 cases were eliminated from the 10th grade sample, including one foreclosure, one 50 8 10 12 Total Sample ONIEIS DIF .684 .319 .634 .552 FOR .700 .684 .758 .722 MOR .200 .481 .677 .517 ACH .362 .537 .637 .553 AAAP STAGE 1 .716 .566 .554 .617 STAGE 2 .703 .756 .634 .705 STAGE 3 .649 .586 .669 .635 STAGE 4 .869 .835 .772 .833 STAGE 5 .751 .697 .821 .769 Figure 8. Reliabilities of the OlVIEIS and AAAP Identity Status D F M A D = identity diffusion F = foreclosure Grade 10 7 7 3 7 M = moratorium A = identity achievement Figure 9. Number of cases in each identity status by grade level, when cases are eliminated for responding in a highly socially desirable way moratorium, two identity achievements, four transition stage scores, and five whose scores were low on all scales. 9 cases were eliminated from the 12th grade sample, including three identity achievements, two transition stage scores, and four whose scores were low on all scales. In all, 29 cases were eliminated by this procedure, including 11 who scored unambiguously in a given identity status, 6 in transition statuses, and 12 who scored low on all scales. When the subjects who responded in the most socially desirable way were eliminated from the data, the results of the AAAP were modified as shown in Figure 10. 51 Developmental Stage 1 2 3 4 5 8 12 2 1 0 l Grade 10 27 1 5 2 1 12 14 3 1 l l n = 72 Figure 10. Number of cases in each developmental stage by grade level when cases are eliminated for responding in a highly socially desirable way Figure 10 reflects the fact that 7 cases were eliminated from the 8th grade sample, including one Stage 1, one Stage 4, two ambiguous multiple stage responders, and three whose scores were below the cutoff on all scales. In the 10th grade sample, 13 cases were eliminated, including 3 Stage 1, 1 Stage 3, 1 Stage 5, 5 ambiguous multiple stage responders, and 3 whose scores were below the cutoff on all scales. In the 12th grade sample, 9 cases were eliminated, including 1 Stage 1, 1 Stage 2, 3 Stage 4, one ambiguous multiple stage responder, and 3 whose scores were below the cutoffs on all scales. Overall, 29 cases were eliminated, including 12 unambiguously assigned to a stage, 8 multiple stage responders, and 9 whose scores were below the cutoffs one all stages. Since the cases eliminated because of high social desirability scores seemed to have little relationship to the hypotheses tested,the research hypotheses were tested both with the full sample and with the sample reduced by the 29 cases which showed high social desirability scores. TESTS OF THE HYPOTHESES The following hypotheses relate to the validation of the OINIEIS with a younger sample. Hypothesis 1 H01: The proportion of students at each grade level in identity diffusion remains the same during early adolescence. HA]: The proportion of students at each grade level in identity diffusion decreases during early adolescence. H01‘ D8=D10=D12 HA1: D8>D10>D12 alpha=.05 The appropriate analysis with the full sample yielded a chi-square value of .308 to be compared to a reference value of 3.914. The null was accepted in this case. The appropriate analysis with the sample reduced because of high social desirability scores yielded a chi-square value of .138 to be compared with a reference value of 3.95. The null was accepted in this case. Hypothesis 2 H02: The proportion of students at each grade level in moratorium remains the same during early adolescence. 53 HA2: The proportion of students at each grade level in moratorium increases with grade level during early adolescence. lV18=M1 0=M12 HA2: M8