Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Relationships between self-esteem, gender role identity, and body image in adolescent girls and their… Usmiani, Sonia 1992

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1992_spring_usmiani_sonia.pdf [ 5.03MB ]
JSON: 831-1.0054187.json
JSON-LD: 831-1.0054187-ld.json
RDF/XML (Pretty): 831-1.0054187-rdf.xml
RDF/JSON: 831-1.0054187-rdf.json
Turtle: 831-1.0054187-turtle.txt
N-Triples: 831-1.0054187-rdf-ntriples.txt
Original Record: 831-1.0054187-source.json
Full Text

Full Text

RELATIONSHIPS BETWEEN SELF-ESTEEM,GENDER ROLE IDENTITY,AND BODY IMAGE IN ADOLESCENT GIRLSAND THEIR MOTHERSbySONIA USMIANIB.A., Dalhousie University, 1980A THESIS SUBMITTEDIN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THEDEGREE OF MASTER OF ARTSinTHE FACULTY OF GRADUATE STUDIES(DEPARTMENT OF COUNSELLING PSYCHOLOGY)We accept this thesis as conformingto the requir^tandardsTHE UNIVERSITY OF BRITISH COLUMBIAMARCH, 1992© Sonia Usmiani 1992Department ofThe University of British ColumbiVancouver, CanadaDate^ 1952 In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature)DE-6 (2/88)ABSTRACTThe present study was conducted in order to investigatethe relationships between self-esteem, gender role identity,and body image in adolescent girls and their mothers. Thestudy included two samples. The first sample consisted ofmenstrual girls and their mothers, while the second samplecontained mother/pre-menstrual daughter pairs. Self-esteemand gender role identity were identified as predictorvariables. Body image was the criterion variable.One hundred and thirteen mother/adolescent daughterpairs volunteered to participate in the study. The BodyImage Subscale of the Self-Image Questionnaire for YoungAdolescents (SIQYA) was utilized as a measure of body image.The Rosenberg Self-Esteem Scale was was utilized as ameasure of self-esteem, and the Bem Sex-Role Inventory(BSRI) was administered in order to measure gender roleidentity. A researcher-designed demographic questionnairewas also administered to all participants. Statisticalanalysis included: 1) step-wise multiple regression analysis2) correlation matrix 3) t-tests on selected variables.Hypothesis I and II stated that for menstrual daughtersand their mothers, self-esteem, as measured by the scores onthe Rosenberg Self-Esteem Scale, and gender role identity,as measured by the Bem Sex-Role Inventory (BSRI) would notpredict body image, as measured by scores on the Body ImageiiSubscale of the Self-Image Questionnaire for youngAdolescents (SIQYA). Significant results were found (on thep < .05 level), and the null hypotheses were thereforerejected.Hypothesis III stated that mothers' body image scoreswould not predict body image scores for their menstrualdaughters. Significant results were found (on the p^.05level), and the null hypothesis was therefore rejected.Hypothesis IV and V stated that for premenstrualdaughters and their mothers, respectively, scores on self-esteem and gender role identity would not predict scores onbody image, as measured by the instruments indicated above.Significant results (on the p^.05 level) were found forpre-menstrual daughters on gender role identity, and onself-esteem for mothers. The null hypotheses were thereforerejected in those cases.A significant correlation was not found between bodyimage scores for premenstrual girls and their mothers. NullHypotheis VI was therefore accepted.iiiTABLE OF CONTENTSABSTRACT^  iACKNOWLEDGEMENTS^ viiCHAPTER ONE  1STATEMENT OF THE PROBLEM^  1Introduction^  1Conceptualization and Definition of Body Image ^ 2Relationship of Body Image to Self-Esteem^ 4Development of Gender Role Identity^ 6Role of Mother/Daughter Relationship  7Impact of Menarche on Adolescent Development... 10Conclusion^  11Purpose of Study^  13Limitations of Study  15Definition of Terms^  16CHAPTER TWO^  19REVIEW OF THE LITERATURE^  19Relationship of Body Image to Self-Concept^ 19Relationship of Body Image to AdolescentDevelopment^  23Relationship of Body Image to Gender RoleIdentity  32Theoretical Underpinnings Underscoring theImportance of the Mother/Daughter Relationship ^ 40Conclusion^  58Statement of Hypotheses^  60ivvCHAPTER THREE^  64METHODOLOGY  64Participants^  64Instrumentation  74Procedure^  83Data Analysis  84CHAPTER FOUR^  87RESULTS  87CHAPTER FIVE^ 105DISCUSSION 105Restatement of the Purpose^ 105Discussion of the Sample 106Summary and Discussion of the Results^ 109Reactions by Participants to the Study^ 124Implications for Future Research^ 127Recommendations for Counselling  129Conclusion^  132REFERENCES^  135APPENDICES  149A. Demographic Questionnaire/Mothers^ 149B. Demographic Questionnaire/Daughters^ 152C. Bem Sex-Role Inventory^  153D. Body Image Subscale of the Self-ImageQuestionnaire for Young Adolescents/Mothers 154E. Body Image Subscale of the Self-ImageQuestionnaire for Young Adolescents/viDaughters^  155F. Rosenberg Self-Esteem Scale^ 156G. Letter of Introduction to Participants^ 157H. Standardized Scatterplot Diagrams^ 158LIST OF TABLES1. Means and standard deviations for demographiccharacteristics of Sample I mothersand daughters^  672. Means and standard deviations for demographiccharacteristics of Sample II mothersand daughters^  673. T-tests on mothers in Sample Iand Sample II  894. T-tests on daughters in Sample Iand Sample II^  905. Means and standard deviations for Sample Imothers and daughters on SIQYA, BSRI, andRosenberg SE Scale^  916. Means and standard deviations for Sample IImothers and daughters on SIQYA, BSRI, andRosenberg SE Scale^  917. Step-wise multiple regressions of SIQYA scoreson independent variables for Sample Idaughters^  938. Step-wise multiple regressions of SIQYA scoreson independent variables for Sample Imothers^  989. Correlation matrix of SIQYA, ROS, and BSRIscores for Samples I and II^  9810.Step-wise multiple regressions of SIQYA scoreson independent variables for Sample IIdaughters^  10211.Step-wise multiple regressions of SIQYA scoreson independent variables for Sample IImothers^  102ACKNOWLEGEMENTS vi iI wish to express my sincere gratitude and appreciation tothe following people:To my family, Renate, Mirko, and Tina, without whoseuntiring support and faith this study would not have beencompleted.To Dr. Judith Daniluk, for her guidance, direction, andinsights.To Bev and Toby, for their patience and humor.Finally, to Lucy, bringer of light.CHAPTER ONEStatement of the ProblemIntroduction Body image has become a topic of increasing interestamong researchers. Characteristically, however, body imagehas been investigated in the context of its relation toeating disorders (Attie & Brooks-Gunn, 1989; Fabian &Thompson, 1989; Mendelson & White, 1985; Strober & Yager,1985; Thompson & Dolce, 1985; Wooley & Wooley, 1985). As aresult, much of our understanding of body image is based onthis collection of literature. An outcome of this researchhas been a trend among researchers of eating disorders tohypothesize that the development of body image hasimplications that reach far beyond this area of inquiry.Contemporary literature reflects an awareness amongresearchers that body image may be related to many aspectsof human development including self-esteem, sexuality,familial relationships, and identity (Attie & Brooks-Gunn,1989; Jackson, Sullivan, & Rostker, 1988; Koff, Rierdan, &Silverstone, 1978; Rosenbaum, 1979).Conceptualization and Definition of Body Image The concept of body image has long been the source ofmuch exploration, analysis, and debate. Theorists in thearea such as Fisher & Cleveland (1968), Schilder (1935), andSchontz, (1969) have struggled with creating a clear1definition of body image. Many definitions have beenoffered, but to date the subject remains the topic of agreat deal of controversy. Recently theorists have arguedthat body image is:the subject as well as the object of is both cognitive and emotional is structure as well as process. The bodyimage has the power to extent beyond the confines ofthe personal body and to engage in social exchangeswith the body images of others (Schontz, 1969, p.170).In a similar vein, Fisher & Cleveland (1968) haveargued that body image functions as a boundary between selfand environment. Theoretically, those individuals withstrong boundaries will be able to function or cope moreeffectively than those with permeable, or diffuseboundaries. There is agreement in the literature that theconcept of body image is multidimensional and complex: itinvolves both internal- biological and psychological factorsand external social factors (Blyth, Simmonds, & Zakin, 1985;Garner, Garfinkel, Schwartz, & Thompson, 1980; Petersen etal., 1984; Rosenbaum, 1979).Schonfield (1964) has argued that an individual's bodyimage is influenced by several factors, including the wayothers react to her, a comparison of her development withthe physical development of those around her, and by a2comparison to cultural ideals. Blyth et al. (1985) note thatthe individual's subjective experience of her body, as wellas her observable physical characteristics, are vital to anunderstanding of body image. An investigation of body imagemust therefore take into account an individual's subjectiveassessment of their bodies. This makes the measurement ofbody image a difficult task. Traditional measures of bodyimage have been criticized on the grounds that they attemptto measure the individual's perception of her size, with noattention to the meanings subjects ascribe to their physicalcharacteristics (Thompson & Dolce, 1989; Wooley & Wooley,1985). A comprehensive measure of body image then, mustinclude and differentiate between cognitive/rational andaffective/emotional aspects (Thompson & Dolce, 1989).In addition, body image is seen as a process which isconstantly in flux: it evolves, is altered and shiftsaccording to life experiences. "Body image is more than anabstract concept; it is a many layered, evolving set ofmemories. It is an integral part of the integrative functionof the mind" (Wooley & Wooley, 1985, p.397). As a result ofthis perspective, body image measures have been developedwhich relate affective components to assessments of bodysize and shape (Petersen et al., 1984). In this study theresearcher attempted to incorporate this broaderconceptualization of body image into the design and overall3structure of the study through the inclusion of a body imagemeasure which taps the affective components of thisvariable.Relation of Body Image to Self EsteemThe results of earlier studies indicate that body imageis an important component of identity development atadolescence (Clifford, 1971; Golombeck, Marton, Stein, &Korenblum, 1987; Koff et al., 1978). Some researchers havecalled for further research which would provide informationon the relation of body image to specific factors such asidentity formation and self-esteem (Attie & Brooks-Gunn,1989; Fabian & Thompson, 1989; Jackson et al., 1988). Forexample, studies investigating the relationship between bodyimage and self-esteem suggest that for women, in particular,there is a tendency for low self-esteem to be correlatedwith negative body image (Jackson, Sullivan, & Rosker, 1988;Jones & Mussen, 1958; Lerner, Orlos, & Knapp, 1976;McCaulay, Mintz, & Glenn, 1978; Noles, Cash, & Winstead,1975; Rosen & Ross, 1968; Secord & Jourard, 1953).This phenomenon is exemplified in a study conducted byMcCaulay et al., (1988). The researchers administered self-report measures to 176 college undergraduates in order toexamine the relationships between body image, self-esteem,and depression-proneness. Results indicated that womenexpressed more dissatisfaction with their bodies than men.4The authors also reported a positive correlation betweenbody image and self-esteem, with lower self-esteem beingassociated with negative body image.Developmental researchers and theorists have arguedthat accommodation to pubertal changes is a keydevelopmental task of adolescence (Attie & Brooks-Gunn,1989; Blyth et al., 1985; Fabian & Thompson, 1989; Petersen,1988; Petersen et al., 1984; Rosenbaum, 1979, Tobin-Richardset al., 1983). While this task must be met by both girls andboys, the impact of physical changes on self-esteem seems tobe greater for girls than for boys. Empirical evidencereveals that adolescent girls tend to report lesssatisfaction with their bodies and correspondingly lowerlevels self-esteem than their male cohorts (Clifford, 1971;Gove & Herb, 1974; Offer et al., 1982; Petersen, 1988;Simons & Rosenberg, 1975; Tobin-Richards, Boner, & Petersen,1983). For example, in an investigation of body image andself-esteem involving 97 adolescent girls and boys,Mendelson & White (1985) administered self-report measuresto determine the relationship of the variables at variousages. The researchers reported that adolescent girls hadsignificantly lower self-esteem and more negative body imagethan their male counterparts.The collection of research on the relationship betweenself-esteem and body image at adolescence demonstrates that5for adolescents, like adults, low self-esteem and negativebody image are correlated, particularly for women and girls.The nature of this relationship, and in particular, itsnegative impact on the self-perceptions of adolescent girlsis as yet poorly understood.Development of Gender Role IdentityResearchers have also identified the formation ofgender role identity as a key developmental task atadolescence (Attie & Brooks-Gunn, 1989; Petersen, 1988;Rosenbaum, 1979). Bem's (1981) gender schema theory suggeststhat the manner in which an individual forms gender roleidentity is a function of the weight they ascribe tocultural definitions of masculinity and femininity. Fromthis perspective, a young girl's self-esteem and self-concept are influenced by the degree to which she believesherself to meet cultural standards. This can be particularlyproblematic for adolescent girls, who may be influenced to agreat degree by contemporary images in the media whichdepict the ideal woman as extremely thin, passive and overlysexualized, while at the same time both career and familyoriented. These conflicting and unrealistic culturalstandards may result in confusion and a sense of inadequacyfor both adult women and adolescent girls (Garner et al.,1980; Petersen, 1988; Steiner-Adair, 1986; Thorbecke &Grotevant, 1982; Wooley & Wooley, 1985).6For example, in their investigation of therelationships between body image and gender role identity,Jackson et al. (1988) administered self-report measures to106 female and 60 male undergraduates. The authors reportedthat women who defined themselves as feminine tended toreport negative body image. In a similar study, Hawkins,Turell, & Jackson (1983) surveyed 80 female and 99 maleundergraduates. Results of self-report measures indicatedthat high scores on femininity were related to negative bodyimage, dieting tendencies, and low achievement motivation,particularly for women.Contemporary research suggests that gender roleidentity is related to body image and self-esteem. Inparticular, evidence from research indicates a tendency forfemininity to be associated with negative body image andself-esteem. The majority of research in this area has beenconducted using adult women. As a result, the relationshipsbetween these variables at adolescence is not understood.Role of Mother-Daughter RelationshipStudies which investigate components of adolescentdevelopment have indicated that familial relationships alsoplay an important role in identity formation (Gold & Yanoff,1985; Golombeck, Marton, Stein, & Koernblum, 1987; Holmbeck& Bale, 1988; Isberg, Hauser, Jacobson, Powers, Noam, Weis-Perry, & Folansbee, 1989; Kamptner, 1988; LeCroy, 1988;7Offer, Ostrov, & Howard, 1982). The relationship betweenmothers and adolescent daughters has been identified as aparticularly significant factor in adolescent development(Holmbeck & Hill, 1986: Isberg et al., 1989; Kamtner, 1988;Leaper, Hauser, Kremen, Powers, Jacobson, Noam, Weis-Perry,& Follansbee, 1989). In particular, the mother-adolescentdaughter relationship has been identified as the strongestpredictor of self-image in adolescent girls (Holmbeck &Hill, 1986; Leaper et al., 1989; LeCroy, 1988; Offer,Astrov, & Howard, 1982). For example, a study conducted byOffer et al., (1982) on 106 teenagers and their familiesreported that the mother-daughter dyad was a strongpredictor of daughters' self-image.These findings are in keeping with both the objectrelations and self in relation theoretical perspectives.According to both these models of development, a girl'sinteractions with her mother form the basic structure, orcontext, within which her self-concept evolves (La Sorsa &Fodor, 1990; Miller, 1976; Surrey, 1985; Rubin, 1983). Thesetheorists posit that the process of development is dynamicand reciprocal; that is, both mother and adolescent daughterare experiencing developmental transitions, and each affectsthe other (La Sorsa & Fodor, 1990; Surrey, 1985; Rubin,1983; Wooley & Wolley, 1985).8Some studies have been conducted on the relationshipsbetween mothers and daughters in terms of gender roleidentity (Jackson, Ialongo, & Stollak, 1986). For example,in their investigation of parental and child gender roles,Jackson et al. (1986) administered self-report measures to184 undergraduates and their parents. The results obtainedthrough a regression analysis of the data revealed thatfemininity scores of the daughters were predicted by thefemininity scores of the mothers.As well, mothers and daughters have beeninvestigated in an effort to identify predictors, or riskfactors in eating disorders. For example, Attie & Brooks-Gunn (1989) conducted a two-year longitudinal study ofmothers and daughters which explored the relationshipsbetween body image, eating disturbance, and familyrelationships in adolescent girls and their mothers. Theauthors reported that body image emerged as a predictor ofeating disturbance for adolescent girls, and that mothers'eating disturbance scores predicted eating disturbance indaughters. Because eating disturbance was defined as thedependent or criterion variable, the study did not make adirect comparison between mothers and daughters on bodyimage scores.Research on adolescent development suggests that forthe adolescent girl, the mother-daughter relationship.may9provide the context in which identity formation occurs(Holmbeck & Hill, 1986; Isberg et al., 1989; Kamptner, 1988;Leaper et al., 1989). In particular, research has indicatedthat the mother-daughter relationship is related to thedaughter's self-esteem, with a close relationship associatedwith high self-esteem (Offer et al., 1982). As well, thereis evidence to suggest that femininity in mothers is relatedto femininity in daughters (Jackson et al., 1986). Existingresearch therefore provides evidence that the mother-daughter relationship may be a critical variable in anyinvestigation of female adolescent development.Impact of Menarche on Adolescent Development There is evidence in the literature that the onset ofmenstruation is a pivotal event in the development of theadolescent girl (Attie & Brooks-Gunn, 1989; Fabian &Thompson, 1989; Koff et al, 1978; Petersen, 1988; Rierdan &Koff, 1980; Rosenbaum, 1978). For example, a study conductedby Rierdan & Koff (1980) involving human figure drawings by153 adolescent girls revealed significant differencesbetween the body images of pre-menstrual and menstrualgirls. The authors reported that the pre-menstrualparticipants emphasized specific body parts, while themenstrual girls demonstrated a more integrated body image.Differences between pre-menstrual and menstrual girlshave also been reported in terms of their relationships with10their families. For example, Hill, Holmbeck, Marlow, Green,& Lynch (1985) utilized data obtained through observationand questionnaires to investigate the associations betweenmenarchal status and child-parent relationships in 62mothers, 61 fathers, and their daughters. The researchersreported results of regression analyses which revealedsignificant differences between girls according to menstrualstatus. For example, mothers were perceived by menarchaldaughters as being less accepting and more controlling thanmothers of premenstrual daughters. As well, menarchaldaughters were reported to participate in family activitiesless often than premenstrual girls. This evidence suggeststhat changes in the mother-daughter relationship may beassociated with menarchal status, and points to theimportance of the onset of menstruation in the identityformation and behavior of the adolescent girl.Conclusion There is support in the literature to suggest thatrelationships exist between individual factors such as self-esteem, gender role identity, and body image in both adultwomen and adolescent girls. For example, as discussedpreviously, studies have been conducted on adults oradolescents which examine relationships between self-esteemand body image, and some studies have been conducted whichexplore the relationships between gender role identity and11body image in adult women (Jackson et al., 1988; Whitely,1983, 1988). However, little research has been conducted onthe relationships between these variables at adolescence. Inparticular, research on gender role identity and itsrelation to body image at puberty has yet to be explored.There is also a paucity of research which explores therelationships between the variables of body image, genderrole identity, and self-esteem in the context of the mother-daughter relationship. Thus, although there is empiricalevidence to suggest that the mother-daughter relationship iscritical to the formation of identity in the adolescent girl(Holmbeck & Hill, 1986; Leaper et al., 1989; Le Croy, 1989;Offer et al., 1982), studies have not been conducted whichattempt to link the variables listed above within thecontext of this relationship.In summary, although research on adolescent developmenthas investigated relationships between the variables of bodyimage, gender role identity, and self-esteem independently,to date investigations have not explored these relationshipswithin a single study. Furthermore, despite evidence offeredby previous researchers that the mother-daughterrelationship is a critical factor in identity formation forthe adolescent girl, studies have yet to explore therelationships between variables in this context. As well,although research indicates that for girls the onset of12menarche acts as a pivotal event for the formation ofidentity, the implications of the changes triggered bymenarche have not been studied in depth. Specifically, therole of the onset of menarche in the adolescent girl'sdevelopment of body image, self-esteem, and gender roleidentity has yet to be examined directly.Theoretical perspectives support the notion that theimpact of the onset of menarche has implications for bothmother and daughter, and triggers an interactive process ofdevelopmental transition (La Sorsa & Fodor, 1990). To date,however, little research has been done to explore the natureof this phenomenon and to add empirical support to existingtheoretical perspectives.Purpose of the Study The purpose of the present study was to explore therelationship between self-esteem, gender role identity andbody image, for mothers and their adolescent daughters.Given the cultural, social, biological, and developmentalsignificance of menarche, the adolescent girls whoparticipated in the study included those who had experiencedthe onset of menstruation within 36 months of the time ofstudy and a similar group of girls who had not yetexperienced menstruation.Body image was identified as the criterion variableunder investigation. Predictor variables were gender role13identity and self-esteem. In the study the researcherattempted to determine a) if body image is related to selfesteem in adolescent girls, b) if body image is related togender role identity in adolescent girls, c) if mothers'ratings of body image, self-esteem, and gender role identityare related to ratings on these variables by theiradolescent daughters, and d) for adolescent girls, ifrelationships between scores on the variables differaccording to menarchal status.Because this investigation was concerned exclusivelywith aspects of the development of body image in women,individuals are referred to as "she" or "her". Women werechosen for the focus of this study due to a large amount ofresearch which suggests that women and girls tend todemonstrate greater dissatisfaction with their bodies, and alower level of self-esteem, than men (Clifford, 1971;Jackson et al., 1988; Petersen, Schulenberg, Abramovitz,Offer, & Jarcho, 1984; Petersen, 1988; Rozin & Fallon,1988). Women and girls were also chosen due to research onthe onset of menarche, which highlights the social andcultural significance of this event in the life of theadolescent girl (Delaney, Lupton, & Toth, 1988).Results of this study may be useful both in terms ofproviding information which may lead to further research,and in the application by clinicians who work in the field14of adolescent counselling. To date, relationshipsbetween the variables under investigation have not beenexamined specifically in the context of the mother-daughterrelationship. Therefore, information provided by this studymay be useful to researchers interested in conducting moreextensive research in this area. As well, results of thisstudy may enhance the understanding of clinicians in thefield of adolescent counselling, including those in theareas of eating disorders, sexuality, and family therapy.For example, an enhanced understanding of the interactivenature of the mother-daughter relationship may aid familyand adolescent counsellors in their approach to thedifficulties experienced by their women clients innegotiating the developmental tasks associated withmenarche. As well, information on the relationships betweenself-esteem, and gender role identity, menarchal status, andbody image may be valuable to both counsellors and educatorsof adolescent girls. This information may be used tointervene when adolescents encounter difficulties in theseareas. For example, an understanding of theinterrelationships among these variables may be useful inthe treatment of eating disorders and body imagedisturbance. As well, the information obtained through thisstudy may be useful to educators in terms of the application15of interventions with younger girls which may act to preventfuture difficulties in coping with developmental tasks.Limitations of the Study The study was limited by the lack of randomization insample selection. Accordingly, the generalizability ofresults was limited by the characteristics of the sampleitself. Thus, it could be argued that the size of the sample(n = 226) and the method of selection (recruitment throughsecondary school students and their parents) resulted in asample which represented only a specific segment of thegeneral population. Inconsistencies in the method ofrecruitment (some participants were recruited throughinteractions with daughters, while others were recruitedthrough interactions with mothers) may have acted as apotentially confounding factor, thus limiting the strengthof the results obtained. The characteristics of the sample,such as a fairly high degree of affluence, predominantlywestern European ethnicity, and non-religiosity limit thegeneralizability of results. As well, selection did notrequire that the participants be members of intact families.Overall, the sample consisted of participants who lived in asetting involving both a father and a mother. As such, theratio of single parent/intact/blended families representedin the sample could have presumably affected the outcome ofthe study. Finally, the focus of the study was on the16relationships between the variables of self-esteem, bodyimage, and gender role identity; conclusions of causationcannot be drawn. This study, therefore, was seen as anexploratory investigation only. Information obtained in thestudy may serve to indicate the direction for futureresearch, and may prove to be valuable for therapeuticapplication by clinicians.Definition of Terms For the purposes of this study, the followingoperational definitions were used:Self-image or self-concept: A multidimensional,phenomenological organization of an individual'sexperiences and ideas about themselves in all aspectsof their lives, manifested through functioning insocial domains and through psychological functioning.(Coombs, 1981; Offer, Ostrov, & Howard, cited inPetersen et al., 1984).Self esteem: an evaluative component of self-image; aunidimensional, generalized sense of self-worth andself-acceptance (Petersen et al., 1984; Rosenberg,1979; Wells & Marwell, 1976; Wylie, 1979).Body image: "the picture of our body which we form in ourmind" (Schilder, 1935, p.11).Gender role identification: the process by which anindividual organizes their self-concept and behavior on17the basis of cultural definitions of masculinity andfemininity. (Bem,1981).Gender role or sex-role: the outcome of gender roleidentification; determined by the degree to whichcultural definitions are integrated into anindividual's self-concept (Bem, 1981).Masculinity: a set of culturally defined prescriptions forattitudes and behavior considered to be appropriate formales. In contemporary Western culture, these includeautonomy, instrumentality, and lack of emotionalexpressivity (Bem, 1981).Femininity: a set of culturally defined prescriptions forbehavior and attitudes considered to be appropriate forfemales. In contemporary Western culture, these includeemotional expressivity, nurturance, and interdependence(Bem, 1981).Androgyny: a term for those individuals who describethemselves as having a high level of both masculine andfeminine traits as conceptually defined for the BemSex-Role Inventory (Bem, 1974).18CHAPTER TWOReview of the LiteratureWhat follows is an outline of literature on theory andresearch in the area of body image. Studies relevant to thestudy being proposed are reported. These include studies onthe relationships between body image and the variables ofself-esteem and gender role identity. As well, relevantresearch and theory on adolescent development as it relatesto body image and the mother-daughter relationship areoutlined. Relevant research and theory are reported underthe following sub-headings: relation of body image to self-concept; relation of body image to adolescent development;relation of body image to gender identity; and theoreticalunderpinnings underscoring the importance of the mother-daughter relationship. Results of previous studies arediscussed in terms of their implications for the focus ofthis investigation.Relation of Body Image to Self-Concept The relationship between body image to self concept hasbeen investigated in order to determine the relationshipsbetween body image and an individual's overall sense of selfworth (Blyth et al., 1985; Petersen et al., 1984; Tobin-Richards et al., 1983). There is a great deal of research tosupport the contention that body image and self esteem are19correlated (Jackson et al., 1988; Jones & Mussen, 1958;Lerner, Orlos, & Knapp, 1976, McCauley, Mintz, & Glenn,1988; Noles, Cash, & Winstead, 1985; Rosen & Ross, 1968;Secord & Jourard, 1953). For example, in a study whichexamined the relation of body image to self-esteem anddepression-proneness in 176 male and female undergraduatestudents, McCauley et al. (1988) reported a strong positivecorrelation between body image and self-esteem in both menand women. In this study, higher levels of satisfaction withbody were associated with higher levels of only onecomponent of self-esteem, identified by the authors associal self-esteem. The researchers concluded that a moreglobal measure of self-esteem would be indicated for futureresearch.This result is supported in a study which compared theassociation between body image and self-esteem in 40 middle-aged women (Van Rackley, Warren, & Bird, 1988). The study,which utilized a global measure of self-esteem, theRosenberg Self-Esteem Scale (Rosenberg, 1965) revealed thatwomen with higher self-esteem also reported higher levels ofsatisfaction with their body image.The focus on body image in women in the investigationconducted by Van Rackley et al. (1988) reflects an interestamong contemporary researchers in the gender differencesthat may exist in men and women's ratings of body image.20Studies have therefore attempted to establish the existenceof such differences, and to examine their extent and nature.For example, Jackson et al.(1988) cite studies whichindicated that compared to men, women evaluate their bodiesless favorably, express more distress and dissatisfactionwith weight, view appearance as more important, and perceivea greater discrepancy between ideal and actual body image(Cash, 1981; Franzoi & Herzog, 1987; Frazier & Lisonbee,1960; Jackson et al., 1987, 1988; Lerner, Karabenick, &Stuart, 1973; Lerner et al., 1976; Noles, Cash, & Winstead,1975; Winstead & Cash, 1984). In their investigation ofgender, gender role identity, and body image, Jackson et al.(1988) reported that for female participants, higher self-esteem was associated with more favorable ratings onphysical appearance, physical fitness, and the importance ofphysical health. While these relationships existed for maleparticipants as well, females viewed physical appearance asmore important than males.A study which compared body image and attitudes towardweight in parents and children (Rozin & Fallon, 1988)revealed distinct gender differences in the responses of menand women. The researchers utilized 97 families, comprisedof 97 mothers and fathers, with 55 daughters and 42 sons.Daughters and sons were college undergraduates. Self-reportmeasures were administered to the sons and daughters and21their families. The authors reported that women demonstratedgreater dissatisfaction with body image than men. As well,the authors found evidence of substantial differencesbetween men and women in a comparison of five measures ofattitudes toward weight and eating, with women reportingmore concern with weight, greater guilt about eating, andmore depression than their male counterparts.While these studies provide evidence for the existenceof a relationship between body image and self-esteem, andidentify women as experiencing lower self-esteem and morenegative body image than men, research evidence does notprovide information on alternate factors which may beassociated with this phenomenon. As well, longitudinalstudies have not been conducted which may provide furtherevidence on the relationships between variables. As aresult, conclusions of causality cannot be drawn based onthe evidence available in current literature.This collection of research indicates that anindividual's sense of self, or self-esteem is related totheir body image. Contemporary research also supports theargument that because women tend to place more importance onappearance, and to report more negative body image than men,their level of self-esteem will tend to be lower, or morenegative, than the level of self-esteem reported by men whoalso report negative body image. Research therefore supports22the contention that the relationship of body image to self-esteem is particularly salient for women. Evidence obtainedthrough research suggests that women tend to experience lowself-esteem and correspondingly negative body image;however, the nature of this relationship is poorlyunderstood. As a consequence, self-esteem is a relevant andcritical component in any investigation of body image inwomen. To that end, the variable of self-esteem was includedin this study.Relation of Body Image to Adolescent Development Body image has been investigated in both women andgirls (Attie & Brooks-Gunn, 1989; Jackson et al., 1988).However, current research has indicated that adolescents inparticular, undergo dramatic changes in their overall self-perceptions, including body image (Petersen et al., 1984;Rosenbaum, 1979; Tobin-Richards, Boner, & Petersen, 1983).The period of adolescence has thus been highlighted by manytheorists as a critical phase in the development of bodyimage.Contemporary researchers have posited that adevelopmental model is appropriate in the study of bodyimage, particularly at adolescence, when pubertal changestrigger "...a psychological state of crisis, a growth crisisnecessitating change" (Rosenbaum, 1979, p.234). A23developmental approach to the study of body image requiresthat body image be studied in the context of the tasksconfronted by individuals during adolescence (Attie &Brooks-Gunn, 1989; Blyth, 1985; Fabian & Thompson, 1989;Petersen et al., 1984; Rosenbaum, 1979; Strober & Yager,1989; Tobin-Richards et al., 1983). Accommodation to therapid physical changes at adolescence has been cited as acritical task of this phase of development (Blyth et al.,1985). Researchers have found that issues of body image andself-esteem are particularly salient at adolescence, whenthe adolescent is struggling to formulate an identity in theface of both physical and environmental changes (Fabian &Thompson, 1989; Petersen et al., 1984; Rosenbaum, 1979;Tobin-Richards et al., 1983).It has been suggested that for both boys and girls, thechanges in physical size and shape that occur at pubertytrigger changes in body image and corresponding changes inlevels of self-esteem (Attie & Brooks-Gunn, 1989; Reirdan &Koff, 1980; Rosenbaum, 1979). There is empirical evidence tosuggest that like adults, adolescents demonstrate acorrelation between self-esteem and body image, with thosewho evaluate their bodies highly reporting higher levels ofself-esteem than those who report negative evaluations oftheir bodies (Jackson et al.,1988; Jones & Mussen, 1958;Lerner et al., 1976; Noles et al., 1985; Rosen & Ross, 1968;24Secord & Jourard, 1953). While these effects exist for bothgirls and boys, sex differences are apparent, with girlsgenerally showing less satisfaction with their bodies andlower levels of self esteem than boys (Blyth et al., 1985;Clifford, 1971; Crockett & Petersen, 1987; Davies & Furham,1986; Gove & Herb, 1974; Offer et al., 1982, Petersen etal., 1984; Simmons & Rosenberg, 1975; Simmons et al., 1973;Tobin-Richards et al., 1983).To date, little research has been done on the effectsof pubertal changes on the adolescent's psychologicalfunctioning. However, research does indicate psychologicalchanges in pubertal adolescents, and suggests that thesechanges are different for boys than for girls. For example,a study conducted by Crockett & Petersen (1987) onadolescent boys and girls revealed that advancing pubertalstatus was associated with enhanced body image and improvedmood for boys, but decreased feelings of attractiveness forgirls. A similar finding was reported by Davies & Furnam(1986), who investigated body satisfaction in 182 girls aged11-18 years. The study provided evidence to suggest thatbody satisfaction tended to decline during adolescence, andreported an association between negative body image and lowself-esteem. Finally, a longitudinal study conducted byBlyth et al. (1985) on body image in 225 girls in earlyadolescence revealed a positive correlation between body25image and self-esteem. Interviews indicated that girls injunior high school reported particularly low self-esteem andnegative body image. The authors speculated that this effectwas due to extra stresses caused by changes in the schoolenvironment.Fabian & Thompson (1989) investigated body image in 61premenstrual and 60 menarchal girls in an attempt toidentify factors which may contribute to eatingdisturbances. Their study explored relationships between thevariables of accuracy of body size estimation, body esteem,self-esteem, depression, eating disturbance, and teasingexperiences regarding appearance. The researchersadministered self-report measures to the participants, andutilized an adjustable light-beam apparatus as a measure ofsize estimation accuracy. Correlational analyses revealedhigh correlations between body esteem (body image), self-esteem, depression, teasing, and eating disturbance. Theauthors therefore suggested that future research shouldexplore these relationships further, particularly in thecontext of their ability to act as risk factors for thedevelopment of eating disorders.For girls, puberty, and specifically the onset ofmenarche, has been pinpointed as a crucial time period inthe development of body image (Attie & Brooks-Gunn, 1989;Fabian & Thompson, 1989; Koff et al., 1978; Rierdan & Koff,261980; Rosenbaum,1979). Adjustment to menstruation requires aprocess of integration of the girl's old self image with hernew image, which in turn may have an impact on her sense ofself: "the process of integrating changes in physicalappearance and bodily feelings requires a reorganization ofthe adolescent's body image and other self representations"(Attie & Brooks-Gunn, 1989, p.71). For the researcher, thisprocess requires attention to biological and cognitivechanges. The impact of menarche, with its implications forsexual identity, can be viewed as a function of theinteraction between situational and ontogenic factors(Petersen et al., 1984). According to this view, therelationships between these factors are crucial, and must beexamined in any investigation of body image at adolescence.Literature on adolescent development points out some ofthe difficulties adolescents typically experience as theyattempt to adjust to pubertal changes. For example, Attieand Brooks-Gunn (1989) note that the onset of puberty causesa rapid accumulation of body fat in girls, precipitatingdissatisfaction with body image, which in turn can result inthe emergence of dieting behavior. Similarly, Steiner-Adair(1986) points out the dilemma girls often confront as theirbodies undergo changes at puberty:On the one hand, adolescence presents girls with thechallenge of coming to terms with their adolescent27bodies; at the same time, society judges girlsaccording to their looks and the culture encouragesgirls to change their body to fit a narrowly definedcultural ideal (p.100).Girls at early adolescence are confronted with the stressesof the onset of puberty with all its accompanyingimplications, and have been demonstrated to show greaterdifficulties in functioning than adolescents at earlier andlater stages of the life cycle (Golombek, Marton, Stein, &Korenblum, 1987).In a three year longitudinal study of 63 adolescents,Golombeck et al. (1987) used a semi-structured interview toexplore seven key personality functions which were derivedfrom object relations theory. The seven functions identifiedwere identity crystallization, maintenance of identity,relatedness, self-esteem, verbal communication, realitytesting regarding self and others, and role assumption. Thetaped interviews were rated according to a PersonalityFunctions Scale (Golombeck et al., 1987). Results of thestudy indicated that early adolescent boys and girlsdemonstrated greater difficulty in all areas of personalityfunctioning as compared to older adolescents. Competence inall areas seemed to increase with age. The authors thereforeidentified early adolescence as a period of high stress, andspeculated that this may be due to both the onset of28puberty, and to the adolescent's confrontation withsignificantly different role expectations.It has been posited that girls at early adolescence aremore preoccupied with body image compared to girls at otherdevelopmental stages; later adolescents having passed theperiod of rapid growth and having therefore been shown todemonstrate a more integrated body image. For example,Rierdan & Koff (1980) investigated body image in a group of153 sixth-, seventh-, and eighth-grade girls throughdrawings of human figures. The authors reported that earlyadolescents are pre-occupied with individual body parts,while later adolescents demonstrate a more integrated andcoherent body image. The researchers suggested that forearly adolescents, changes in body parts have significancebecause they signal the onset of puberty and sexualmaturity. They posit that later adolescents have experiencedthe physical changes, and are in the process of integratingthese changes into their identities.In a similar study which also utilized drawings ofhuman figures as a measure of adolescent girls' body image,Koff, Rierdan, & Silverstone (1978) reported significantdifferences in the responses between premenstrual andmenarchal participants. This investigation, which involved87 seventh grade girls, reported that menarchal girlsdemonstrated a better articulated and defined body image29than the premenstrual girls. The authors therefore suggestedthat menarche is a pivotal event for the adolescent girl,requiring radical reorganization of body image. Thisresearch suggests that puberty, and the physical changesthat occur as a result, have complex implications foradolescent girls' body image and self-esteem. As a result oftheir research, Rierdan & Koff (1980) and Koff et al. (1978)suggest that investigations of body image in adolescentgirls must address differences among girls as a function ofpubertal and menarchal status.Delaney, Lupton, & Toth (1988) discuss our culture'sdistaste and denial of menstruation: " rather than celebratethe coming-of-age in America, we hide the fact of menarche,just as we are advised to deodorize, sanitize, and removethe evidence" (p.107). In interviews with parents andchildren, the authors discovered that girls were reluctantto discuss the topic of menstruation, and when they werewilling to discuss their feelings regarding this topic,their comments indicated a sense of distaste and shame.Their investigation revealed that the implications ofmenstruation for sexuality and childbirth seemed lessimportant to adolescent girls than the physical discomfortand menarche's influence on social life. They acknowledgedthe influence mothers have on their daughters' attitudestoward menstruation, but recognized that the negative30messages received through the media tend to have a far morepowerful effect. Nonetheless, in their conclusion, Delaneyet al. (1988) commend the efforts of several mother-daughterteams who are in the process of writing books and pamphletsfor adolescents. The authors thereby emphasize theimportance of the mother-daughter relationship, and statethat society's attitudes towards menarche have acted as abarrier between mothers and daughters. For example, untilrecently, literature on menstruation has been dominated bymanufacturers of sanitary products because, the authorsstate, "...ordinary women were too firmly in the grip oftheir society's taboos and their own ignorance to reach outcollectively to their daughters" (p. 114).This collection of research indicates that self-esteemand body image tend to become more negative at earlyadolescence compared to earlier and later phases ofdevelopment. It has been hypothesized that this drop inself-esteem is due to the stressors of this period, whichinclude entry into junior high school, the onset of puberty,the beginnings of heterosexual concerns, and the beginningof dating (Petersen et al., 1984).In summary, literature on body image at adolescenceindicates that self-esteem and body image are related; likeadults, adolescents tend to report higher levels of self-esteem when they evaluate greater satisfaction with their31bodies (Jackson et al., 1988; Noles et al., 1985). As well,research has offered evidence that gender differences existin body image and self-esteem. This research suggests thatgirls tend to report lower levels of self-esteem and lesssatisfaction with their bodies than boys (Petersen et al.,1984; Tobin-Richards et al., 1983). Finally, puberty, andfor girls the onset of menarche has been pinpointed as acritical phase in the development of body image (Attie &Brooks-Gunn, 1989; Koff et al., 1978; Rierdan & Koff, 1980;Rosenbaum, 1979). Thus, research supports the argument thatthe relationship of body image to self esteem isparticularly relevant for girls who are experiencingpubertal changes. In this study the researcher willtherefore examine the relationship between self-esteem andbody image in young girls in the time period surrounding theonset of menstruation.Relation of Body Image to Gender Role Identity The onset of menstruation can be viewed as adevelopmental marker: the adolescent girl experienceschanges that signal the transition from girlhood towomanhood, and must begin to identify and integrate themeanings such change have for her. It has been suggestedthat the formation of identity at adolescence includes thecrucial task of integrating physical changes with societaldefinitions and expectations (Attie & Brooks-Gunn, 1989;32Rosenbaum, 1979). This is a continuation of the process ofgender role identification begun in early childhood. Genderrole, the outcome of this process, is determined by thedegree to which cultural definitions are integrated into anindividual's self-concept (Bem, 1981). Bem has describedthis phenomenon as "the process by whichsociety...transmutes male and female into masculine andfeminine" (p.354).Bem's (1981) gender schema theory is based on acognitive-developmental model. Her theory proposes that sex-typing arises from schematic processing in whichunderstanding is developed through the organization of newinformation based on pre-existing schema, or previousunderstanding. This process involves the assimilation ofself-concept into gender-schema, which is then evaluatedbased on social definitions:the child learns to evaluate his or her adequacy as aperson in terms of gender schema, to match his or herpreferences, attitudes, behaviors, and personalattributes against the prototypes stored within it. Thegender schema becomes a prescriptive standard orguide...and self-esteem becomes its hostage (p.355).Bem (1981) discusses the implications of this processin terms of the differences individuals demonstrate in theextent to which they utilize cultural definitions in their33assessments of their gender role identities. The authorstresses that it is the process itself which is central toher theory, rather than the content of the gender schema:Accordingly, sex-typed individuals are seen asdiffering from each other not primarily in terms of howmuch masculinity and femininity they possess, but interms of whether or not their self-concepts andbehaviors are organized on the basis of gender (p.356).According to this theory, the manner in which anadolescent girl's gender role identity develops will be afunction of the importance or weight she ascribes tocultural definitions. Accordingly her self-concept and selfesteem will be influenced by the degree to which herphysical and personal characteristics meet societalstandards. Individuals will vary in their adherence tocultural norms; those who utilize cultural definitions to ahigh degree in their formation of gender role identity willpresumably be influenced by these norms more than those whorely more heavily on personal or psychological factors. Itis arguable, therefore, that an investigation into bodyimage and self-esteem should include an exploration ofgender role identity.Literature on the relationship of gender role identityand self concept supports this argument ( Jackson et al.,1988; Whitely, 1988). For example, it has been proposed that34gender and gender role are related; that is, women tend toreport feminine traits as being more self descriptive thanmasculine traits, and men are more likely to identifymasculine traits as being self-descriptive than are women(Jackson et al., 1988).As well, research has indicated that masculine andandrogynous men and women demonstrate higher levels of selfesteem than their more feminine counterparts (Jackson etal.,1988; Whitely, 1983, 1988). The term "androgynous" isused to describe those who report high levels of bothmasculine and feminine traits.For example, in an investigation of the relationshipsbetween gender, gender role identity, and body image,Jackson et al. (1988) administered self-reportquestionnaires to 166 undergraduate students (106 femalesand 60 males). The authors hypothesized that women who ratethemselves as being feminine would have a more negative bodyimage than their more masculine or androgynous peers, andwould also consider their appearance to be more importantthan any other group. The researchers reported that femininefemales demonstrated a more negative body image than womenwho rated themselves as masculine or androgynous. However,there were no differences among women in the importance theyascribed to physical appearance. In their discussion ofresults, the authors cited this finding as evidence of the35acceptance among contemporary women of cultural standardsabout the importance of appearance for women. They addedthat because the relationship between gender role identityand body image may be complex and mediated by a variety ofother characteristics of the individual, future researchshould take a broader approach to include a variety ofpsychological and behavioral measures.In a related study of 205 introductory psychologystudents, Whitley (1988) administered self report measuresof self-esteem, including measures of traits and behavior,as well as sex role orientation. Data were analyzed using amulti-trait, multi-method matrix. The results indicated thatgender role identity is related to self esteem, and linksfeminine gender role identity with lower self-worth.There is also evidence to support the contention thatgender role identity and self-esteem are related to bodyimage. Research has been reported which suggests that formen, masculinity is related to body image satisfactionratings, while for women, femininity is correlated withdissatisfaction with body image and concern with weight(Hawkins et al., 1983; Jackson et al., 1987, 1988; Kimlickaet al., 1983; Winstead & Cash, 1984). For example, Hawkinset al. (1983) administered self-report questionnaires to atotal sample of 361 undergraduates in order to investigatethe relationship of weight concerns to gender role identity.36They reported that for women, femininity was positivelycorrelated with dieting tendencies, and that dissatisfactionwith body, negative body image, was associated with lowself-esteem.Based on existing research, it is arguable that womenwho are highly gender typed, that is, women who describethemselves as having traits typically associated withfemininity, are more likely to report negative body imageand low self-esteem. While there is a paucity of researchwhich investigates the relationship of gender role identityand self-esteem in adolescents, it is reasonable tospeculate that the relationships evident in the research onadults may also exist for adolescents. Evidence fromresearch on adolescent development which emphasizes theimportance of social influences and relationships inidentity formation, particularly for girls (Thorbecke &Grotevant, 1982) adds credence to this argument.Although the majority of research on gender roleidentity has been conducted with adult participants, theliterature on adolescent development offers both theoreticaland empirical perspectives on the significance this variableat adolescence. For example, research on adolescencesuggests that the process of establishing gender roleidentity is a critical component of overall identityformation (Golombeck, Marton, Stein & Korenblum, 1987;37Jackson, Ialongo, & Stollak, 1986; Kamptner, 1988;Rosenbaum, 1979). Developmental theorists such as Erickson(1968) have proposed that at adolescence, a key task is theformation of an overall sense of identity, which consists ofa variety of components. In her discussion of femaleadolescent development, Rosenbaum (1979) utilizes theframework offered by Blos (1967) to identify four majorchallenges or tasks of adolescence: idividuation andseparation, ego-continuity, coping with residual trauma, andsexual identity. The author describes sexual identity as theprocess by which the adolescent comes to integrate maturesexual functioning into the sense of being female or male.The physical changes that signal sexual maturation thustrigger a process of structuring social definitions offemininity or masculinity into overall identity. Resolutionof this task will result in the establishment of gender roleidentity.Research on adolescent development has identifiedgender differences between adolescent girls and boys interms of the ways in which they address issues of identity(Kamptner, 1988; Jackson et al., 1986; Thorbecke &Grotevant, 1982). The component of gender role identity, inparticular, has been identified as an area of differencebetween boys and girls. For example, in a study of 41 maleand 42 female adolescents involving interviews and self-38report measures, Thorbecke & Grotevant (1982) reported thefinding of an interrelationship between occupational andinterpersonal domains for girls. That is, for girls,interpersonal variables,such as self-esteem, were found tobe related to occupational interests and aspirations. Asimilar finding was not revealed for boys. The authors positthat this result may be due to the conflicts in gender roleswhich exist for contemporary girls. That is, theestablishment of gender role identity is problematic forgirls because they are currently being socialized formarriage and motherhood while simultaneously beingencouraged to seek career goals. The authors posit thatbecause traditional social definitions of femininity arebeing questioned in the absence of the formulation of newdefinitions, the acquisition of gender role identity isparticularly difficult for the contemporary adolescent girl.In summary, this collection of research providesfurther evidence for the argument that for women, self-esteem and body image appear to be related, with low self-esteem being related to negative body image (Jackson et al.,1987, 1988; Kimlicka et al., 1983; Winstead & Cash, 1984).This research also provides evidence which suggests a linkbetween gender role identity and body image by reportingthat for women, perceptions of femininity tend to be relatedto negative body image (Jackson et al., 1987, 1988; Kimlicka39et al., 1983; Winstead & Cash, 1984). Based on researchcited previously (Jackson et al., 1987, 1988; Kimlicka etal., 1983; Winstead & Cash, 1984) it would seem reasonableto argue that the variable of gender role identity may be animportant component of body image, particularly for girls atpuberty. In this study therefore, relationships betweenself-esteem, body image, and gender role identity wereexamined in adolescent girls in the months preceeding andfollowing the onset of menarche.Theoretical Underpinnings Underscoring the Importance of the Mother-Daughter Relationship Research on adolescent development has reinforced theimportance of family relationships in the formation ofidentity for adolescent girls (Holmbeck & Hill, 1986; Isberget al., 1989; Kamtner, 1988; Leaper, Hauser, Kremen, Powers,Jacobson, Noam, Weis-Perry, & Follansbee, 1989). It has beenreported that parenting styles characterized by warmth andacceptance serve to enhance the identity formation processat adolescence for both boys and girls (Coopersmith, 1967;Gecas, 1972; Isberg, Hauser, Jacobson, Powers, Noam, Weis-Perry, & Follansbee, 1989; Kamptner, 1988; Macoby & Martin,1983; Rosenberg, 1965). However, gender differences in boththe nature of the relationship, such as the level ofintimacy, and its impact on identity have been noted.Researchers have identified the relationship between mothers40and daughters as being the strongest predictor of the self-esteem of adolescent girls (Holmbeck & Hill, 1986; Leaper etal., 1989; LeCroy, 1988; Offer et al., 1982). For example,in an investigation into the impact of parent-adolescentintimacy on adolescent functioning involving 85 adolescents(LeCroy, 1988), analysis of data obtained through self-report measures revealed that mothers received higherratings on intimacy than fathers, and the level of parentalintimacy was positively correlated to adolescent self-esteem, particularly for girls.In a related study, Holmbeck & Hill (1986) employed apath-analytic model to investigate the relationships betweenparent-reported instrumental and expressive traits, child-reported parental acceptance, and adolescent self-esteem andself-consciousness in 159 sixth and seventh graders andtheir families. The researchers identified the mother-daughter dyad as the relationship in which the strongestcorrelations occurred. They reported that for mothers anddaughters, maternal expressiveness was predictive of child-reported maternal acceptance; acceptance, in turn wasreported to be predictive of self-esteem for daughters.A similar result was reported by Offer et al. (1982) intheir study of family perceptions of adolescent self-image.The researchers administered self-report measures to 241adolescents and their parents in an effort to examine the41relationships between adolescents' self-image and theperceptions of the adolescents held by their parents. Theauthors reported that mother-child congruence of perceptionwas more highly correlated with adolescent self-image thanfather-child congruence. In addition, the study revealedthat daughters' self-image was more highly associated withparent-child agreement than was sons'.In a study which explored family communicationpatterns, Leaper et al.(1989) examined the psychosocialchanges occurring at adolescence from the perspective ofgender role socialization. This study involved 32 adolescentboys and girls aged 14-15 years from two parent middle- andupper-class families. Participants and their parents wereasked to engage in a discussion of moral dilemmas. Theirresponses were then coded using the Constraining andEnabling Coding System. The results of the study indicatedpatterns of communication which suggested that separationfrom parents occurred more often with conformist sons andnon-conformist daughters, while closeness with parentsoccurred more often with conformist daughters andnonconformist sons. The authors posit that developmentalchanges in relationships are different for females andmales, with emphasis on autonomy for boys and intimacy forgirls. "The challenge of psychosocial development can thusbe viewed as a search for both individuality and communion,42to balance one's striving for autonomy and one's need forinterpersonal connections" (Leaper et al., p.336). Theresults of the study suggest that parents treat sons anddaughters differently. For example, unlike interactionsbetween fathers and sons or mothers and sons, mother-daughter communications are characterized by closeness andintimacy. Traditional or sex-typed maternal communicationswere found to predict daughters who were considered to be inthe conformist level of Loevinger's (1976) model of egodevelopment. This level is characterized by "...a groupcentered orientation in which external rules and socialacceptability are guiding principles" (Leaper et al.,p.338). Results of this study suggest that the nature ofparent-child interactions may have a significant impact onthe child's attitudes and behavior. As well, resultsindicate that traditionally gender-typed mothers tend tohave daughters who are similarly invested in adherence tosocietal standards of appropriate gender-based behavior.In a study which investigated the relationships betweenparents' gender role identities and those of their children,Jackson et al. (1986) surveyed 184 undergraduates and theirparents using self-report measures. A regression analysis ofthe data identified maternal femininity as a strongpredictor of femininity for daughters. This study did notinclude adolescent participants; however, it is possible43that these results could be extrapolated to a youngerpopulation. Further research is therefore indicated in thisarea.The nature of the mother-daughter relationship itselfhas also come under investigation. For example, daughterswho perceive mothers as affectionate and accepting have beenreported to score higher on measures of self-esteem(Holbeck & Hill, 1986; Offer et al., 1982). In a studyinvolving 67 married women aged 28-38 years, Sholomskas andAxelrod (1986) found that women who perceived theirrelationships with their mothers as loving and acceptingscored higher on self-esteem and role satisfaction measures.Further, there is evidence that daughters who reporthigh levels of affection and identification with theirmothers show a corresponding high level of intimacy withpeers (Gold & Yanof, 1985). In this study the researchersadministered questionnaires to 134 high school girls in aneffort to examine the relationship between the mother-daughter relationship and peer relationships. Theresearchers reported a significant relationship betweenmother-daughter relationships characterized by affection,identification, and democratic treatment and peerrelationships characterized by intimacy and mutuality.There is also evidence to suggest that the mother-daughter relationship undergoes changes according to the44daughter's menarchal status. For example, Hill et al.,(1985) identified the period directly following the onset ofmenarche as a time of heightened stress between mothers anddaughters. This investigation analyzed data obtained throughobservation and questionnaires from 62 mother-daughter and61 father-daughter dyads. Results indicated that daughtersreported the perception that mothers were less acceptingthan they were before menstruation began. Menarchaldaughters also reported that they were less influenced bymothers, and did not participate in family events asfrequently. It is arguable that for girls, relationship withmother may be a primary context in which developmental tasksare undertaken, and this relationship may therefore be apowerful factor in the daughter's personal development.Contemporary developmental theory supports the notionthat the mother-daughter relationship is a critical factorin the formation of gender identity for an adolescent girl(Miller, 1976; Rubin, 1983). Current literature onadolescent development reflects differences amongdevelopmental theorists in their conceptualizations ofdevelopmental processes. For example, traditionaldevelopmental theorists have proposed that the tasks ofadolescent development include cognitive shifts fromconcrete operations to utilization of reflective thought andpropositional logic, increasing physical and psychological45autonomy, and the evolution of a stable and cohesive self-structure (Strober & Yager, 1989). This view of developmentas an evolution toward greater separateness and autonomy hasbeen criticized by feminist theorists on the grounds that itdoes not take into account the experiences and life viewsunique to women (Gilligan, 1982; Miller, 1976, 1983;Steiner-Adair, 1986; Surrey,1985; Rubin, 1983).In her discussion of a developmental model for women,Miller (1983) posits that the models put forward bytheorists such as Erikson (1950) and Levinson (1978), withtheir emphasis on development of self through separation-individuation do not describe women's experiencesadequately. She argues that for women, development of selfoccurs in the context of relatedness to others. According tothis theoretical perspective, interactions with others,particularly mother, are hypothesized to form the structurewithin which a woman's self concept evolves. Thus, it iswithin the context of relationships that a woman developsher sense of herself.Her various roles such as mother,friend, co-worker, or daughter and the relationships sheexperiences within these roles, provide her with criticalinformation on her identity, and allow her to locate herselfwithin her environment.Rubin (1983) examines differences in the process ofidentity formation between boys and girls from the46perspective offered by object relations theory. According tothis theoretical perspective, the mother-offspringrelationship is critical to development, as it forms thefoundation on which self concept and gender identity arebuilt. The mother-child relationship is identified as beingcritical because, in most societies it is a woman, and mostoften the mother who acts as the child's primary caregiver.Rubin argues that for boys, a sense of gender identityarises from the recognition as a small child that he isdifferent from mother, and must separate from her andidentify himself with father if he is to form and maintainan independent sense of self. This process requires that heset rigid boundaries between himself and mother. Once theseboundaries are in place, they serve to facilitate hismovement toward greater autonomy, and set the tone for allfuture relationships .For girls however, this process of gender role identityformation through separation is not considered necessary,because mother is not different from daughter (Rubin, 1983).Therefore, the process of gender role identity formation ismore ambiguous; the young girl need not wrench herself fromher relationship with mother, and as a result, does notdevelop the rigid ego boundaries characteristic of youngboys. Rubin discusses the implications of this pattern ofdevelopment for women:47The context within which separation takes place andidentity is forged means a girl never has to separateherself as completely and irrevocably as a boy must.Her sense of herself, therefore, is never as separateas his; she experiences herself always as morecontinuous with another; and the maintenance of closepersonal connections will continue to be one of life'sessential themes for her. (pp. 58-59)For a girl then, relationship with mother is viewed as beingcritical for two reasons. First, it is critical because itserves initially to shape her sense of gender role identityand her manner of relating to others. Second, because a girldoes not separate completely from her mother, her process ofidentity formation at adolescence will be a continuation ofthe process begun in early childhood, whereby interactionwith mother serves as the context within which herdevelopment occurs.Proponents of the self in relation model of developmentecho this theoretical perspective (Miller, 1983; Surrey,1985). These theorists argue that for an adolescent girl,self worth develops in the context of her relationships withfamily and peers: "the girl's sense of self-esteem is basedin feeling that she is a part of relationships and is takingcare of the relationships" (Miller, 1983, p.5). For a woman,the struggle at adolescence may not be toward greater48autonomy; instead, the adolescent girl attempts to retainher sense of herself as "being-in-relationship" (Miller,1983, p.9) without losing her sense of agency within herrelationships. Miller (1983) emphasizes that the younggirl's understanding of "being-in-relationship" does notimply dependency; instead, relationship involves mutualcaretaking, flexibility, and growth.In her discussion of the nature of the mother-daughterrelationship, Surrey (1985) posits that through herrelationship with her mother, a girl develops theexpectation that the mutual sharing of experiences will leadto psychological growth, and thus that psychologicalconnection will enhance the development of her self concept.Therefore, for both mothers and daughters, "self esteem...isrelated to the degree of emotional sharing, openness, and ashared sense of understanding and regard" (p.6). Because therelationship between mother and daughter involves mutualsharing and openness, the process of development of identitythrough connection is considered to be reciprocal. Thus, theself concepts of mothers, as well as daughters, arereciprocally affected.According to this model of development, both mother anddaughter create and are affected by the quality of theirrelationship. Thus, a daughter's body image may be affectedby her mother's attitude toward her body. However, because49the relationship is reciprocal, a mother's body image may beequally affected by the attitudes of her daughter. Surrey(1985) notes that:The ability to move and change in relationship clearlydepends on the capacities and willingness of all peopleinvolved to change and grow, not just the child. Sincethis growth is interactional, it is often difficult tosee who "leads" or "initiates" this process (p.8).This theoretical perspective is valuable as a context toinvestigate the relationships of body image for mothers anddaughters, within the context of a reciprocal and dynamicprocess.Reinforcing the concept of reciprocity, there isevidence to suggest that the dynamics of the mother-adolescent daughter relatio, .ip are intensified as themother copes with both her own identity in mid-life, as wellas her changing role as a mother of an increasingly adultdaughter. For example, in a study of parent-adolescentinteractions, Silverberg & Steinberg (1987) interviewed andsurveyed individual members of 129 intact families. Theirresults demonstrated that mothers report more intensemidlife identity concerns and lower self-esteem when theyhave daughters who are more physically mature. This researchsuggests that both mothers and daughters are affected by thedaughter's developmental status.50La Sorsa & Fodor (1990) support this position, arguingthat both mother and adolescent daughter are experiencing adevelopmental crisis of separation and self-definition:Separation and self-definition are two of the maindevelopmental tasks for the adolescent daughter as wellas for her mid-life mother. The specific dynamics ofthe mother-daughter relationship serve to complicatethe achievement of these developmental tasks. The factthat the adolescent daughter and the mid-life motherare both confronting these challenges simultaneouslycan either enhance or diminish each other's attemptstowards mastery and growth. (p. 600)The authors argue that the period of adolescence appears tointensify the mother/daughter relationship, and may requireshifts in self-definition for both the daughter and themidlife mother.According to LaSorsa & Fodor (1990), the period of thedaughter's adolescence may involve a parallel process ofreassessment of both individual identity and therelationship itself. For the daughter, puberty and theaccompanying physical changes triggers a process of renewedidentity formation; issues of sexuality, independence,responsibility, and career must be addressed. For themother, the daughter's development has implications formany, if not all, aspects of her life. For example, midlife51signals physical changes, including weight gain, wrinkledskin, increased health concerns and often, menopause. Thebeginning of the daughter's menstrual cycle may thereforesignal the end of the mother's. The mother must thenstruggle with the issues surrounding her own sexuality, aswell as that of her daughter's. As well, the daughter'sincreasing independence often allows her mother greaterfreedom; time to reassess career goals, perhaps leading tothe return to school or changes in her current job status.Finally, because this is an interactional process, thedevelopmental processes of mother and daughter appear to beinfluenced, and to a certain extent, mediated by each other.The relationship itself must be redefined; issues ofseparation, boundaries, competition, and rebellion arise(La Sorsa & Fodor, 1990). The authors argue that in orderto address these issues, contemporary theory and researchrequires a dyadic interactional model, with a cognitive-developmental perspective.In addition, as discussed previously, there is evidenceto suggest that mothers who demonstrate traditional gender-typed characteristics have daughters who place a high valueon societal standards (Leaper et al., 1989). It would seemreasonable to assume, therefore, that there may be aninteraction between a mother's self-esteem and gender role52identity and the self-esteem and gender role identity of herdaughter.As well as focussing on gender role identity,contemporary research reflects an interest in other aspectsof sociocultural influences on body image. For example,Western society's current standard of thinness, along withits emphasis on youthfulness, is often included indiscussions of the etiology of body image dissatisfactionamong women, particularly in relation to eating disorders.(Attie & Brooks-Gunn, 1989; Garner et al, 1980; Steiner-Adair, 1986; Wooley & Wooley, 1985).Wooley & Wooley (1985) link cultural influences tonegative body image, and discuss the implications ofsocietal standards for mothers and daughters. The authorsargue that because cultural standards are transmittedthrough the family, a mother's distress caused by herfailure to meet impossible cultural standards will affecther daughter's body image. For example, these authorscontend that a mother's preoccupation with her body andeating behavior may have a profound effect on herinteractions with her child. They conclude that "...the mostprobable assumption is that the process of feeding is markedby inconsistency and is greatly emotionally charged"(p.399). The authors cite a study conducted by Debs et al.(1983) as an illustration of their position. This study, in53which college women were surveyed, found that the strongestpredictor for bulimia (after self-evaluation), was thedaughters' belief that her mother was critical of herdaughter's body.In a related study, Steiner-Adair (1986) tested herhypothesis that eating problems were tied to femaleadolescents' perceptions of cultural values that make itdifficult for girls to integrate and value relationships asthey develop. The researcher utilized self-report measuresand a clinical interview in her investigation of 32 girlsaged 14-18. The results of this study indicate a tendencyfor women who accept cultural standards without question toassociate thinness with autonomy, success, and recognitionfor achievement. Steiner-Adair argues that the perception ofa connection between autonomy, success, and thinness placesgirls and women at increased risk of eating disorders:It is a vision of autonomy and independence whichexcludes connection to others and a reflectiverelationship to oneself...In this light, it is possiblethat eating disorders emerge at adolescence because itis in this point of development when females experiencethemselves to be at a crossroads in their lives wherethey must shift from a relational approach to life toan autonomous one, a shift that can represent an54intolerable loss when independence is associated withisolation. ( p.107)Wooley & Wooley (1985) posit that dissatisfaction withbody may lead to dieting by the daughter, which the authorsidentify as a common precursor to eating disorders. Theauthors contend that dieting by the daughter may serve toenhance her identification with her mother, whilesimultaneously differentiating herself by assuming controlover her body.In a two-year longitudinal study Attie & Brooks-Gunn(1989) investigated maternal body image, eating patterns andattitudes, depression, and family systems and relationshipsas possible predictors of eating problems in adolescentgirls. Their study involved the administration of self-report measures to 193 adolescent girls and their mothers.Participants were contacted two years after initial contact,and questionnaires were re-administered. Regression analysesrevealed that of the variables under investigation, familycharacteristics was the only significant predictor ofdaughters' eating concerns (9%). Mothers' scores on eatingpatterns approached significance as a predictor ofdaughters' eating concerns (p < .10). Maternal body imagewas not identified as a significant predictor of daughters'eating difficulties. The study did not attempt to comparedaughters' body image scores directly with those of their55mothers. It did, however, identify daughters' body image asthe most powerful predictor of eating problems.This position is supported in a study conducted byRozin and Fallon (1988). These researchers administeredquestionnaires to 97 families, related to body imagesatisfaction and concern with weight. Results indicated thatsignificant differences appeared between opposite-sex pairs(ie., father-mother, son-daughter), while mothers anddaughters reported similar concerns. In their comparison ofeating attitudes and body image among mothers, fathers,sons, and daughters, Rozin & Fallon found that mothers anddaughters resembled each other in their concerns about bodyimage and eating, while fathers, although dissatisfied withweight, did not show the same concern or distress. Theresearchers concluded that gender is a much better predictorof body image attitudes than is generation. Although notconclusive, this study provides further evidence of females'greater reported dissatisfaction with body than males', andsuggests that these patterns exist within the context of,and may be perpetuated by mother-daughter interactions.In summary, literature on female adolescent developmenthas identified the mother-daughter relationship as a keycomponent in the development of daughters' identity, and inparticular, self-esteem (Holmbeck & Hill, 1986; Leaper etal., 1989; LeCroy, 1988; Offer et al., 1982). Research has56also identified puberty, and specifically, the onset ofmenarche as being pivotal for the daughter's identityformation (Hill et al., 1985; Koff et al., 1978; Petersen,1988; Rierdan & Koff, 1980). Physical changes that accompanypuberty and the onset of menstruation may trigger a processof re-definition of identity for both mother and daughterand necessitate the restructuring of the relationship itself(La Sorsa & Fodor, 1990). Therefore, research has identifiedthe mother-daughter relationship as a primary context inwhich developmental tasks are addressed (Hill et al., 1985;LaSorsa & Fodor, 1990; Silverberg & Steinberg, 1987; Wooley& Wooley, 1985).The concept of reciprocity between mother and daughteris supported by object relations theory ( Rubin, 1983).According to this theoretical model, identity formation atadolescence is a continuation of a process begun in infancy,wherein infants identify themselves in terms of theirrelationships with their mothers. For girls, puberty, andspecifically the onset of menarche may cause anintensification of this process. Issues of gender roleidentity and body image may become particularly salient asthe girl becomes increasingly physically mature. As well,daughters' physical maturity may have implications for bothmother and daugh:er, who must then address issues ofseparation and individuation (Attie & Brooks-Gunn, 1989;57Hill et al., 1985; LaSorsa, 1990; Thorbecke & Grotevant,1982). Finally, gender role identity, self-esteem and bodyimage in mothers and daughters may be influenced or mediatedby societal definitions of femininity; definitions which arecurrently in flux (Rozin & Fallon, 1988; Steiner-Adair,1986; Wooley & Wooley, 1985). As such, the responses of bothmothers and their adolescent daughters on the measures ofself-esteem, gender role identity, and body image wereincluded in this investigation.Conclusion Literature on both adult women and adolescent girlssuggests a relationship between body image and self-esteem(Attie & Brooks-Gunn, 1989; Jackson et al., 1988; Petersenet al., 1984; Tobin-Richards et al., 1983). Studiesinvestigating gender differences in body image suggest thatfor women, there is a tendency for low self-esteem to berelated to negative body image (Cash, 1981; Franzoi &Herzog, 1987; Frazier & Lisonbee, 1960; Jackson et al.,1987, 1988; Lerner et al., 1973, 1976; Noles et al., 1975;Winstead & Cash, 1984). Because studies of these variableshave been correlational in nature, conclusions of causalitycannot be drawn.Research on adolescent development has indicated thatfor girls at puberty, the relationship of body image toself-esteem may be particularly salient, with a tendency for58low self-esteem to be associated with negative body image(Attie & Brooks-Gunn, 1989; Petersen, 1988; Rosenbaum,1979). The onset of menarche has been identified as beingpivotal for the adolescent girl, in terms of overallidentity formation, and specifically for body image (Attie &Brooks-Gunn, 1989; Fabian & Thompson, 1989; Koff et al.,1978; Rierdan & Koff, 1980). Investigations of body image atadolescence should therefore focus on the time perioddirectly surrounding the onset of menstruation so that theeffects of this variable may be better understood.Research on adolescent development also suggests thatthe onset of menarche triggers a process of theestablishment of gender role identity, a key component ofoverall identity (Bem, 1981; Hill et al., 1985; La Sorsa &Fodor, 1990). Theory on adolescent development suggests thatthe onset of menarche has critical implications for theadolescent girl, as it signals her transition from girlhoodinto mature womanhood. As a result, issues of socialdefinitions of masculinity and femininity must be addressedand integrated into overall identity (Rubin, 1983; La Sorsa& Fodor, 1990).Literature on adolescent development therefore suggeststhat issues of self-esteem, body image and gender roleidentity are critical variables in identity formation. In59addition, for girls, the onset of menarche has beenidentified as a pivotal event in this process.Further, because theory and research suggest that themother-daughter relationship may be the primary context fora girl's development, it is arguable that this relationshipmay be a critical variable in understanding therelationships between body image, self-esteem, and genderrole identity (Hill et al., 1985; La Sorsa & Fodor, 1990;Miller, 1976; Rubin, 1983). It has also been suggested thatboth mothers and daughters are affected by socio-culturalinfluences such as gender role prescriptions and culturalstandards for thinness (Rozin & Fallon, 1988; Wooley &Wooley, 1985). The impact of societal prescriptions forappearance and behavior on adolescent body image shouldtherefore be investigated in the context of the reciprocaldynamic interactions between mothers and daughters.In order to address these issues, variables of bodyimage, self-esteem, and gender role identity were examinedin adolescent girls and their mothers. This approach madepossible the independent examination of relationshipsbetween variables for mothers and daughters, as well as aninvestigation of the associations between mothers anddaughters on these relationships.60Statement of the Hypothesis In attempting to provide clinicians and theorists inthe field of adolescent development with information on therelationships between the criterion variable, body image,and the predictor variables, gender role identity and self-esteem, the following hypotheses were generated.Hypothesis I A) Menstrual daughters' self-esteem scores on the RosenbergSelf-Esteem Scale do not significantly predict daughters'body image, as measured by the Body Image subscale of theSelf-Image Questionnaire for Young Adolescents (SIQYA).B) Menstrual daughters' gender role identity scores on theBem Sex-Role Inventory (BSRI) do not significantly predictdaughters' body image, as measured by the SIQYA.Hypothesis II A) The self-esteem scores of mothers of menstrual daughterson the Rosenberg Self-Esteem Scale do not significantlypredict mothers' body image scores, as measured by theSIQYA.B) The gender role identity scores of mothers of menstrualdaughters on the Bem Sex-Role Inventory (BSRI) do not61significantly predict mothers' body image scores as measuredby the SIQYA.Hypothesis III For mother/menstrual daughter pairs, mothers' body imagescores on the Body Image subscale of the Self-ImageQuestionnaire for Young Adolescents (SIQYA) do notsignificantly predict daughters' body image scores asmeasured by the SIQYA.Hypothesis IVA) Pre-menstrual daughters' self-esteem scores on theRosenberg Self-Esteem Scale do not predict daughters' bodyimage scores as measured by the SIQYA.B) Pre-menstrual daughters gender role identity scores onthe Bem Sex-Role Inventory (BSRI) do not significantlypredict daughters' body image scores as measured by theSIQYA.Hypothesis VA) The scores of mothers of pre-menstrual daughters on theRosenberg Self-Esteem Scale do not significantly predictmothers' body image scores as measured by the SIQYA.62B) The gender role identity scores for mothers of pre-menstrual daughters on the Bem Sex-Role Inventory (BSRI) donot significantly predict body image scores as measured bythe SIQYA.Hypothesis VI For mother/pre-menstrual daughters, mothers' body imagescores on the Body Image Subscale of the Self-ImageQuestionnaire for Young Adolescents (SIQYA) do notsignificantly predict daughters' body image scores asmeasured by the SIQYA.63CHAPTER THREEMethodologyParticipants A total sample of 113 mother/daughter pairs, or 226participants was utilized in the present study. A total of442 questionnaire packages were distributed. One hundred andtwenty nine packages were returned to the researcher. Ofthese, 16 were eliminated from the study. The majority ofthese (n = 11) were eliminated because only one member ofthe mother-daughter pair completed the questionnaires. Twopackages were returned with incomplete responses from bothmother and daughter, and three were returned long after thedata had been analyzed. The response rate was thereforecalculated to be 28.5 %.The total sample was divided into two samples: a firstsample of 82 mother/daughter pairs, which consisted ofmother/daughter pairs in which the daughters had begunmenstruating within the previous 36 months, and a secondsample of 32 mother/daughter pairs, in which the daughterswere pre-menstrual. The first sample were referred to asSample I and the second sample, Sample II. Criteria forinclusion of daughters were as follows:1) Girls aged 11-14 years were targeted because themean age of onset of menstruation in Western countriesis 12.5 years ( Tapley, Morris, Roland, Weiss, Subak-Sharpe, & Goetz, 1989). Sample 1 girls were requested64to have experienced their first menstruation in the 36months preceding the study, to ensure that they wereexperiencing pubertal changes, and correspondingchanges in body image at the time of the study. Theperiod of 36 months was specified in order to provide arange of time after onset of menarche for menstrualperiods to be experienced on a regular monthly basis.2) it was required that all participants were not underpsychiatric care, and that they were not under theinfluence of mood altering medication, to alleviatepotential distortion of the data caused by medication,or psychiatric illness.3) it was required that the family had not undergonetransition or trauma caused by divorce, major illness,or death of a family member at the time of study, toensure that responses were not likely to be influencedby other traumatic life circumstances.Mothers were required to meet criteria numbers 2 and 3.Object relations theory places emphasis on the nature of therelationship between mother and daughter beginning ininfancy and continuing through adolescence ( Rubin, 1983).Mothers were therefore required to have lived with thedaughter since birth, in order to ensure that the sample washomogeneous in terms of early family constellations.Consistent with object relations theory, it was viewed as65important that mothers and daughters had co-habitated forthe majority of the child's life. The study therefore didnot require participant mothers to be the biological mothersof the daughter participants, and no requirement was madethat the mother-daughter pair be living in an intact familysetting.Demographic information on the characteristics of thesamples was obtained through the use of two experimenter-generated "Demographic Questionnaires". The first wasdesigned for mothers (see Appendix A) and the latter fordaughters (see Appendix B). Information on the mothers anddaughters in Sample I will be reported first (see Table 1),followed by information on mothers and daughters in SampleII (see Table 2).Sample IMothers:Mothers in Sample I (N = 82) ranged in age from 32 to55 years, with a mean age of 42.7 years (SD = 4.31). Heightfor mothers in both samples ranged from 59 to 70 inches,with a mean height of 64.75 (SD = 2.17), while weight rangedfrom 95 to 250 pounds, with a mean weight of 136.4 pounds(SD = 24.3).The educational level of the mothers who participated66TABLE 1Sample IMeans and Standard Deviations for DemographicCharacteristics of Mothers and DaughtersMothers DaughtersM^SD M SDage^(years) 42.7 4.31 13.4 1.12height^(inches) 64.75 2.17 64.5 2.63weight^(pounds) 136.44 24.26 114.5 14.82grade 8.64 1.01age of 1st. 12.09 .897menstruation^(years)n = 82 pairsTable 2Sample IIMeans and Standard Deviations for DemographicCharacteristics of Mothers and DaughtersMothers DaughtersM^SD M SDage^(years) 44.86 4.98 12.52 .962height^(inches) 65.16 2.54 61.52 4.45weight^(pounds) 136.26 19.48 99.69 11.84grade 7.80 .792age of 1st.menstruation (years)n = 31 pairs67in the study was relatively high, with only 1.2% of thesample reported having completed elementary school, 36.6%completing high school, 22.0% completing college, 24.4%completing university, and 14.6% completing graduatestudies. In terms of employment status, the majority ofmothers (70.7%) were active in the work force. Of this70.7%, 32.9% worked full-time and 37.8% worked part-time. Inaddition, 24.4% of mothers reported their primary occupationas homemaker, and 4.9% reported being unemployed. The mostfrequently reported occupations were in the fields of health(15%), and education (15%). Participants reported a widediversity of occupations, including areas such as business,design, foreign service, and creative pursuits such aswriting. Of the total sample, 57.3% of mothers reportedhaving been employed in the first five years of theirdaughters' lives, while another 41.5% reported that theywere not employed during that time. Of the total sample,1.2% of mothers did not respond to this question.When asked to report family income, 70.7% of mothersclassified their families as having a total income whichexceeded $46,000 per year. Another 12.2% placed their totalfamily income in the $36,000 - $45,000 per year category. Ofthe total sample, only 6.1% reported their family income asbeing less than $25,000 per year, with most women in thiscategory being single parents. Eighty six percent of mothers68reported that a father lived in the home, while 13.4%reported that a father did not live in the home. However,participants were not asked to specify if the family wasintact or blended. Stepdaughters who were not included inthe study were reported by 1.2% of mothers, and stepsons byan additional 1.2%. Daughters included in the study were notreported to be step-daughters by their mothers.Twenty-nine percent of mothers claimed that theyparticipated in an organized religion on a regular basis,while 69.5% did not participate. Religious affiliation wasreported as follows: 53.7% Christian, 40.2% Agnostic, 3.7%Hindu, 1.2% Muslim and 1.2% Jewish. Frequency ofparticipation ranged from at least 4 times per month (17.1%)to never (51.2%). Participation in an organized religion ata rate of 1-2 times per month was reported by 12.2% ofmothers, while 3.7% reported participating 3-6 times peryear, and 15.9% reported participating 1-2 times per year.The most frequently reported ethnic heritage formothers was classified as Western European (56.1%). Thisclassification included European countries such as England,Holland, Sweden, France, and Italy. For daughters the ethnicclassification reported most frequently was Western European(45.1%).When asked whether their family situation had changedin the two to three years preceding the study, 80.5% of69mothers responded in the negative, while 19.5% respondedthat there had been changes. When asked to describe thetypes of changes that had occurred, responses varied fromdeath of grandparents (8.5%) to addition of a family member,such as the birth of a child (1.2%) or the arrival of anephew to live in the home (1.2%). Divorce, separation,and/or re-marriage was reported by 8.5% of the respondents.Because these respondents stipulated that the changes hadoccured beyond the time limits established in the criteriafor inclusion in the study, these participants weretherefore included in the study.Daughters:Daughters in Sample I (N = 82) ranged in age from 11 to16 years, with a mean age of 13.42 years (SD = 1.12). Schoolgrade level ranged from grade 7 to grade 11, with a meangrade of 8.65 (SD = 1.01). Height ranged from 56 to 70inches with a mean of 64.5 (SD = 2.6), and weight rangedfrom 82 to 150 pounds, with a mean of 114.5 pounds (SD =14.8).Daughters' religious orientation (as reported by themother) was similar to their mothers'. For example, 32.9% ofdaughters were reported to participate in a religiousorganization on a regular basis, while 67.1% did not. Aswell, daughters' religious affiliation was reflective ofthat of their mothers' with 50% reporting Christian, 43.9%70Agnostic, 3.7% Hindu, and 1.2% each for Muslim and Jewish.Frequency of participation for daughters was only slightlyhigher than for mothers. For example, 22.0% of daughterswere reported to attend a religious organization at leastfour times per month, as compared to 17.1% for mothers.For daughters in Sample 1, the age of onset ofmenstruation ranged from 10 to 15 years, with a mean of 12.1(SD = .89). Daughters in this group reported to have beenmenstruating for a period of time that ranged from 6 to 36months, with a mean of 9.46 months.Sample IIMothers Mothers in Sample II ( N = 31) had a mean age of 44.86years (SD = 4.98). Age ranged from 36 to 57 years. Meanheight for this sample was 65.16 inches (SD = 2.54) with arange of 60 to 70 inches, and mean weight was 136.26 pounds(SD = 19.48), with a range of 95 to 190 pounds. Educationallevel was reported to range from high school (32.3%) tograduate studies (25.8%). Approximately one quarter ofmothers reported to have completed college (25.8%), and16.1% completed university.Current employment status was reported as follows:29.0% full-time, 32.3% part-time, 32.3% homemaker, and 3.2unemployed. A small percentage (3.2%) did not respond tothis question. Of those who were employed outside the home,7122.6% described their employment to be in the field ofeducation, 19.4% in the field of health, 9.7% in sales, and3.2% in business. An additional 9.7% of respondents workedin a variety of fields, including various types ofadministration, design, the arts, and manual labour. Themajority of mothers (71.0%) reported that they did not workin the first five years of their daughters' lives.Family income was reported as follows. The majority ofrespondents (77.4%) reported income to exceed $50,000 peryear, while 9.7% reported income to be less than $25,000 peryear. Three percent of respondents placed their incomes inthe categories of $26,000-$35,000 per year and #36,000-$50,000 per year.Like mothers in Sample I, most mothers in Sample IIreported that a father lived in the home (83.9%). Step-daughters were not reported in this group, and only 3.2%reported step-sons.Approximately one quarter of respondents (25.8%)reported that their family situation had changed in theyears preceding the study. However, the changes describeddid not warrant exclusion from the study.The majority of mothers in Sample II (58.1%) reportedthat they did not participate in an organized religion on aregular basis. Of those who claimed a religious affiliation,48.8% described themselves as Christian, 32.3% as Agnostic,729.7% as Jewish, with an additional 9.7% reporting variousaffiliations including Fundamentalist and New Age. Frequencyof participation in an organized religion was reported asfollows: 4 times per month, 25.8%, 1-2 times per month,12.9%, 1-2 times per year, 16.1%, and never, 45.2%.Like mothers in Sample I, the ethnic heritage mostcommonly reported by mothers in Sample II was WesternEuropean (45.2%). An additional 38.7% reported theirethnicity as Canadian, 3.2% as East European, 3.2% asAmerican, and 9.7% as Hebrew.Daughters:Daughters in Sample II (N = 31) had a mean age of 12.5years (SD = .962), with a range of 11 to 14 years. Heightranged from 52 to 69 inches, with a mean height of 61.5inches (SD = 4.45). Weight for this group ranged from 73 to132 pounds, with a mean weight of 99.69 pounds (SD =11.84). Mean school grade was 7.8 (SD = .792), and rangedfrom grade 7 to grade 9 (see Table 2).A minority of daughters were reported by their mothersto participate in an organized religion (38.7%). Daughters'religious affiliation was as follows: Christian, 48.4%,Agnostic, 35.5%, and Jewish, 9.7%. An additional 6.5%reported a variety of affiliations which included New Age,Pentecostal, and Quaker. Frequency of participation in73organized religion ranged from 4 times per month (25.8%) tonever (51.6%). Nine and seven-tenths percent of daughterswere reported by their mothers to participate 1-2 times permonth, and 12.9% were reported to participate 1-2 times peryear.Daughters' ethnic heritage was predominantly Canadian(41.9%) or West European (41.9%). A small number of mothersreported their daughters' ethnic heritage to be EastEuropean (3.2%), American (3.2%), and Hebrew (9.7%).for this variable, most notably in Sample II. Scores ongender role identity were moderate for all participants, andfell within the range of scores obtained by the appropriatenorm groups (Bem, 1981).Instrumentation For the purpose of this study, four questionnaires wereadministered to all participants. The Body Image Subscale ofthe Self-Image Questionnaire for Young Adolescents( Petersen, 1984) was utilized as a measure of body image.The Rosenberg Self-Esteem Scale (Rosenberg, 1965) provided ameasure of self-esteem. Gender role identity was determinedthrough the administration of the Bem Sex-Role Inventory(Bem, 1981). Finally, two experimenter-generatedquestionnaires (designed separately for mothers anddaughters) were utilized in order to obtain demographicinformation on the participants.741) Bem Sex-Role Inventory  (BSRI) (Bem, 1981). The BSRI isthe most frequently used of all gender-type scales(Lippa, 1985). The BSRI is a self-administered measurewhich asks participants to rate themselves onstereotypically "feminine" or "masculine" traits. Items arerated on a 7-point Likert-type scale ranging from "never oralmost never true" to "always or almost always true" (seeAppendix C). The measure provides 3 scores: femininity,masculinity, and femininity-minus-masculinity difference.Since its introduction, the original form of the BSRI (Bem,1974) has been modified to a "short form" which contains 30items, half its original number (Lippa). Modifications weremade in order to address criticisms aimed at the originalform, and have resulted in a scale which is considered to bepsychometrically superior and factorially purer than itspredecessor (Lippa). As well, the short form scalescorrelate strongly (.90) with corresponding scales of theoriginal BSRI (Lippa), and with the Personal AttributesQuestionnaire (PAQ) (Holmbeck & Bale, 1988).Lippa (1985) reports the following coefficient alphasfor internal consistency and reliability for the BSRI-SF:for females, .75 for the femininity scale and .87 for themasculinity scale; for males .78 for femininity and .87 formasculinity. Normative data are provided in the BSRI manualfor non-Stanford black, white, and Hispanic undergraduates,75psychiatric patients, and groups of different age categoriesincluding youth, young adults, and older adults.Scores from the BSRI short form are given in the formof T-scores based on the difference between scores on thefemininity and masculinity scales. The femininity-minus-masculinity scores range from - 50 to + 50, with a score of0 separating masculinity on the negative side to femininityon the positive side. T-scores are calculated based on thisscore. The resulting scores range from 12, which representsgreatest masculinity to 88 which represents the highestscore obtainable for femininity.Evidence has been presented for the construct validityof the BSRI: for example, Holmbeck and Bale (1988) conducteda multitrait, multi-method analysis which demonstratedconvergent and discriminant validity of the instrument. Theauthors reported that for males, convergent validitycoefficients were .35 and .62 when r (.0001) = .25. Forfemales, convergent validity coefficients were .52 and .30when r (0001) = .17. Discriminant validity coefficients forthe BSRI were reported to be -.27 for males when r(.001)=.20 and .15 for females when r (.001) = .12 . Aswell, results of a study conducted by Frable (1989) arepresented as evidence for the predictive validity of theBSRI. This study, in which the relation of sex-typing togender ideology was examined, reported results of a76multivariate analysis of variance (MANOVA) which show asignificant main effect of 2.41 when p < .005. These resultsprovide evidence for the predictive validity of the BSRI bydemonstrating its ability to predict relevant constructssuch as gender role identity at a statistically significantlevel.The BSRI has been used extensively in research ongender role orientation. For example, it has been used instudies in which the relation of gender role orientation toself-esteem was investigated both in adults (Kimlicka,Cross, & Tarhai, 1983; Moore & Rosenthal, 1980; Sethi &Bala, 1983) and adolescents (Lamke, 1982). Studies whichcompared gender role orientation between males and femaleshave included the BSRI (Etaugh & Weber, 1982; Kimlicka,Wakefield, & Goad, 1982; Korabik, 1982). As well , the BSRIhas been administered in studies in which gender roleidentity and social and familial relationships wereinvestigated (Falbo, 1977; Feldman, Biringen, & Nash, 1981;Fleck, Fuller, Malin, Miller, & Acheson, 1980). Because theBSRI is an established measure of gender role orientationwith a high number of available norms, it was used withinthe present study as a predictor measure of gender roleidentity. The brevity of this measure allowed the researcherto obtain information in an efficient manner.772) Self Image Questionnaire for Young Adolescents (SIQYA)(Petersen et al., 1984). This measure of self-image is adownward extension of the Offer Self-Image Questionnaire(Offer, Ostrov, & Howard, 1982) and utilizes nine of itsscales (see Appendix D & E). These scales include EmotionalTone, Impulse Control, Body Image, Peer Relationships,Family Relationships, Mastery and Coping, Vocational andEducational goals, Psychopathology, and Superior Adjustment.The instrument consists of 98 items. The number of items foreach subscale ranges from 8 to 17. Responses are elicited ona 6-point Likert-type scale, ranging from (1) "describes mevery well" to (6) "does not describe me at all". Items arescored so that a high score indicates high self-image(Petersen et al., 1984). Psychometric properties of thescale have been demonstrated using three samples: a studysample (n=335), a longitudinal sample (n=253), and avalidating sample (n=343). Subjects were both male andfemale white upper middle class sixth grade students (meanage 11.6), who were followed through to the eighth grade.Petersen et al. (1984) offer evidence for the validityof the SIQYA by stating that it is highly correlated with awell-validated measure of self-esteem, the Rosenberg Self-Esteem Inventory. Construct validity was ascertained by theadministration of both the SIQYA and the Rosenberg Self-Esteem Inventory (SEI) to 335 seventh grade students in a78single testing session. Correlations for overall self imagewere reported to be .72 for boys and .62 for girls. As well,members of the same sample (n=335) were screened to findevidence of mental health problems. A multivariate analysisof variance demonstrated that those students who reportedmental health problems also report poorer self-image overalland poorer self-image on appropriate scales on the SIQYAthan the nonproblem group.For the purposes of this study, the Body Image Scale,an 11 item subscale of the SIQYA, was utilized as a measureof mothers' and daughters' body image satisfaction (seeAppendix D and E). This scale taps affective and socialcomparative aspects of body image (Petersen et al., 1984).Internal reliability of this scale is reported in the formof alpha coefficients for interitem consistency (Petersen etal.). The authors report that for the Body Image Scale, thecoefficient for reliability is .81 for boys and .77 forgirls. The high rating reported for internal reliability(Brooks-Gunn, Rock, & Warren, 1989; Petersen et al., 1984)ensures that the subscale can be used in isolation from theother subscales without compromising the validity of theresults. Petersen et al. argue that an advantage of theirinstrument over other measures of self-esteem is that theresults with specific scale scores, body image, inparticular, demonstrate differentiated, and therefore79informative patterns. The body image subscale is thereforehighlighted by its authors as having the capacity to provideaccurate information independently of the other subscalescomprising the SIQYA.Scores for the Body Image Subscale of the SIQYA arecalculated based on raw scores, with high scoresrepresenting positive body image. Scores range from 1.0 to6.0, which represents the highest possible score obtainable.The Body Image Subscale of the SIQYA has beenadministered to mothers and daughters in previous research.For example, Attie and Brooks-Gunn (1989) utilized this 11item scale in their longitudinal study on the relation ofmaternal characteristics to the development of eatingdisorders in a sample of 193 white adolescent girls aged 13years at first contact. The authors report that slightmodifications were made to the scale for mothers.Modifications were made in order to increase face validity,and were considered by the authors to be insufficient toinvalidate results. For the purposes of this study, similarmodifications were also necessary on items # 34 , 70, and 72of the Body Image subscale. In its original form, item #34read "I am uncomfortable with the way my body isdeveloping". For mothers, this item was changed to read "Iam uncomfortable with the way my body has developed". Item #70 was modified from "When others look at me they must think80that I am poorly developed" to "When others look at me, theymust think that my body did not develop well". Finally, formothers item #72 was modified from its original form of "Mybody is growing as quickly as I would like it to" to "Mybody developed the way I would have liked it to". Anadvantage of this scale for mothers and daughters is that itprovides information on the mother's current body image, aswell as her body image as she developed at adolescence. Inthis way, the scale may provide a link between mothers anddaughters in terms of their perceptions of their experiencesat puberty.3) Rosenberg Self-Esteem Scale (Rosenberg, 1965). This scalewas originally developed for use with high school students,and is designed to measure the self-acceptance aspect ofself-esteem (Rosenberg, 1965). The scale consists of 10items answered on a 4-point Likert-type scale (see AppendixF). Responses range from "Strongly Agree" to "StronglyDisagree". The initial norm group was a sample of 5,024junior and high school seniors from 10 randomly selected NewYork schools (Rosenberg, 1965). Scores are analyzed based onraw scores, which may range from 1.0, representing thelowest score for self-esteem, to 4.0 which represents thehighest level of self-esteem obtainable.Reliability has been reported in the form of a Guttmanscale reproducibility coefficient of .92 (Robinson & Shaver,811973). Test-retest reliability is reported as .85 (Silber &Tippett, 1965). Correlations offered as evidence for thescale's convergent validity include .59 - .60 whencorrelated with the Coopersmith Self-Esteem Inventory, and.27 when correlated with the CPI self-acceptance scale(Robinson & Shaver, 1973). Evidence for the scale'sdiscriminant validity is cited in the form of correlationswith measures of self-stability of .21 to .53 (Robinson &Shaver, 1973). The scale has been widely used since itsinception and is considered by many researchers to be abrief and thorough measure of self-esteem (Robinson &Shaver, 1973). Also, because it was originally developed andnormed using an adolescent population, it is applicable tothe adolescent group under investigation in the presentstudy.4) Demographic Questionnaires (experimenter-generated). Twodemographic questionnaires were designed by the researcherin order to obtain information on the characteristics of thepopulation under investigation. The first was designed formothers, and requested basic information such as height,weight, and age (see Appendix A). Additional information,including occupational status (both current and in the firstfive years of the daughter's life), family income level, andfamily membership was also requested. Mothers were alsorequested to identify religious affiliation and extent of82participation for themselves and their daughters. Thedemographic questionnaire also requested information onmothers' and daughters' ethnic heritage.The second demographic questionnaire was designed forcompletion by daughters (see Appendix B). This questionnairerequested basic information only. Information requested fromdaughters included age, height, weight, and grade. As well,daughters were asked to identify the age and gender ofsiblings, and the age at which they had experienced theirfirst menstrual period. Finally, a section was madeavailable to daughters to express comments and reactions tothe study.Procedure Principals of private schools, directors of schoolboards, and heads of Parent Associations in the GreaterVancouver area were contacted both by telephone and inwriting as a means of recruiting participants. Contact wasmade, and permission to recruit was obtained from HillsideSchool and York House School. Physical Education classesconsisting of girls in grades 7-11 were targeted forrecruitment. Girls in these classes were offered anexplanation of the study and the requirements and parametersof their participation, and questionnaire packages weredistributed to those interested. Students were asked to takethe packages home to be completed once their mothers'83consent and agreement to participate had been obtained. Aswell, Parent Associations from various secondary schools onthe West Side of Vancouver agreed to allow the researcher toenter their meetings in order to introduce the study toparents, and to give interested parents the opportunity toparticipate with the cooperation of their daughters.Questionnaire packages included a letter ofintroduction (Appendix G) which explained the nature of thestudy and offered the names and telephone numbers of theresearchers in the event participants had any questions orconcerns regarding their participation. In addition to theinstruments and the demographic questionnaire, theparticipants were offered a note of thanks, and were giventhe opportunity to request the results of the study as soonas they became available. Finally, the questionnairepackages included a self-addressed, stamped envelope so thatthe completed questionnaires could be returned to theresearcher.Data Analysis Analysis of the data was undertaken in two stages. Inthe first stage, a stepwise multiple regression analysis wasconducted to test the stated hypotheses. Body image, asmeasured by the Body Image Subscale of the Self-ImageQuestionnaire for Young Adolescents (SIQYA) was thecriterion variable. Predictor variables were self-esteem, as84measured by the Rosenberg Self-Esteem Scale, and gender roleidentity, as measured by the Bem Sex-Role Inventory (BSRI).The Statistical Package for the Social Sciences (SPSS) wasutilized for all statistical analyses.Multiple regression analysis was the statisticalprocedure of choice because the study was non-experimentalin nature. As well, this technique was most appropriatebecause the design required that measures on the predictorvariables be obtained in order to assess their ability topredict the criterion variable of body image. Following themultiple regression analysis, there was some exploratoryanalysis to examine the independent contributions of thevariables. Included in the exploratory analysis was anexamination of the similarity dimension of the scores ofmothers and daughters through an analysis of relationshipsbetween their scores on the variables under investigation.Specifically, a series of t-tests were conducted in order todetermine the relationships between various demographicvariables such as religious orientation and the criterionvariable, body image. Finally, t-tests were undetaken whichcompared Samples I and II on demographic variables. A t-testcomparing a subsample of same-aged daughters from eachsample was conducted in order to investigate differences inscores between menstrual and pre-menstrual girls.85In the second stage of analysis, a correlation matrixof mothers versus daughters using body image scores wasproduced in order to assess if there was a relationshipbetween individual mothers and their daughters in theirscores on the predictor and criterion variables. Thecorrelational matrix provided the researcher with specificinformation on body image scores for distinctmother/daughter pairs.86CHAPTER FOURResultsResults were obtained for two samples: Sample I,containing mother/daughter pairs in which the daughter hadexperienced the onset of menstruation in the 36 monthspreceding the study (N = 82 pairs) and Sample II, containingmother/daughter pairs in which the daughter had not yetexperienced her first menstruation (N = 31 pairs). Thehypotheses were accepted or rejected on a p < .05 level ofsignificance.Comparison of Samples on Demographic Information A t-test was conducted to determine whether there weredifferences in demographic characteristics between mothersin Sample I and Sample II. Results of t-tests are presentedin Table 3. Differences were not found on most demographiccharacteristics, with the exception of age (T = -2.19, DF =110, 2-tail prob.= .031). It was therefore determined thatmothers in Sample I were significantly younger than mothersin Sample II. As well, significant differences were foundbetween mothers in Samples I and II on employment during thefirst 5 years of the daughters' lives (T = -2.82, DF = 110,2-tail prob. = .006). T-tests therefore revealed thatmothers in Sample I were employed more often than mothers inSample II during that time.87T-tests on differences between daughters revealedsignificant differences in age between daughters in SamplesI and II (T = 3.99, DF = 111, 2-tail prob. = .0001), height(T = 4.18, DF = 102, 2-tail prob. = .0001), weight (T =4.84, DF = 104, 2-tail prob. = .0001), and grade (T = 4.16,DF = 111, 2-tail prob. = .0001) ( see Table 4). T-teststherefore revealed that girls in Sample I were older,taller, heavier, and in a higher grade than theircounterparts in Sample II.Scores for Dependent and Independent Variables Scores for all participants on the variables underinvestigation are presented in Table 5 for Sample 1 andTable 6 for Sample II. In general, scores on self-esteemwere high, as mean scores were above 2.00 out of a possiblescore of 4.00. Mothers' scores on self-esteem were higherthan daughters' scores on this variable for both samples.Scores on body image were also fairly high for allparticipants, as the mean scores for all groups were above3.00 out of a possible score of 6.00. Again, mothers' scoreswere higher than daughters' scores8889TABLE 3T-tests on Mothers in Sample I and Sample IIVariablesProbilityT DF 2-Tailage -2.19 110 0.031*height -0.85 109 0.397weight 0.04 111 0.970education -.086 110 0.393current employmentstatus^(yes/no) -0.77 110 0.445employed in 1st 5 years ofdaughter's life^(yes/no)^-2.82 110 0.006*family income -0.12 106 0.906father in home^(yes/no)^-0.37 111 0.715n = 113p^. 05TABLE 4T-tests on Daughters in Sample I and Sample IIVariables T DF 2-Tail Probabilityage 3.99 111 0.0001*height 4.18 102 0.0001*weight 4.84 104 0.0001*grade 4.16 111 0.0001*n = 113p^.0590TABLE 5Sample IMeans and standard deviations for mothers anddaughters on SIQYA, BSRI, And Rosenberg SE ScaleMothers DaughtersM SD M SDSIQYABSRIROS4.077.52.943.12.98811.65.4804.03352.952.94.89411.12.522n = 82 pairsTABLE 6Sample IIMeans and Standard Deviations for Mothers andDaughters on SIQYA, BSRI, & Rosenberg SE ScaleMothers^ DaughtersM^SD^M^SDSIQYA 4.57 .713 3.93 .912BSRI 51.71 9.79 51.77 8.77ROS 3.22 .461 2.96 .523n = 31 pairs91To test Hypothesis I (A), a step-wise multiple regressionanalysis was conducted on the scores obtained by menstrualdaughters on the three instruments utilized in the study.Scores for Sample I on the three measures including meansand standard deviations for both mothers and daughters arepresented in Table 5. Results of the first regressionanalysis are presented in Table 7. For daughters in SampleI, step-wise multiple regression analysis revealed that thevariable of self-esteem contributed independent andsignificant proportions of variance to the prediction ofbody image (33.3%). For menstrual daughters, high self-esteem scores were associated with positive body imagescores. Because the independent variable of self-esteem wasfound to make a significant contribution to body imagescores at the^.05 level, the null hypothesis wasrejected, indicating that self-esteem scores were found topredict body image scores at a significat level.To test Hypothesis I (B), a step-wise regressionanalysis revealed that the predictor variables gender roleidentity and self-esteem contributed 37.3% of the variancein body image scores for menstrual daughters. Because self-esteem was determined to account for 33.3% of the variance,it was concluded that gender role contributed an additional4.0 % of the variance in body image scores. An F testconducted on the change in R squared when gender role92TABLE 7Sample IDaughtersStep-wise Multiple Regressions of SIQYA Scores onIndependent VariablesIndependent^RegressionVariables Equation R^21R Square R Square PROS^Y l = 1.29+.99x^.577^.333^ .01BSRI^Yl= 1.99-.02x^.611^.373^.04^.01n = 82 pairsp^.0593identity was added was significant at the p < .01 level,F(2,79) = 5.02, p^.01. Therefore, results indicated thathigh masculinity was associated with positive body image.Gender role identity contributed to variance in body imagescores to a lesser degree than self-esteem. Consequently,the relationship evident between masulinity and positivebody image was considered to be a significant but lesspowerful phenomenon. Because the independent variable ofgender role identity contributed to a significant degree tothe variance in body image scores at the p^.05 level, thenull hypothesis was rejected, indicating that scores ongender role identity were found to predict body image scoresat a significant level.To test Hypothesis II (A), a step-wise multipleregression analysis was conducted using scores obtained frommothers on the three instruments utilized in the study.Results of the regression analysis for mothers are presentedin Table 8. It should be noted that cautions in theinterpretation of the results of regression analyses formothers of menstrual daughters must be taken based on thefinding of multicolinearity, or correlations between thepredictor variables (see Table 9). This result is addressedin the discussion of the testing of Hypothesis III tofollow.94TABLE 8Sample IMothersStep-Wise Multiple Regressions on SIQYA Scores onIndependent VariablesIndependent^ReressionVariables Equation^R AR Square R Square PROS^Y4=.87+1.00x^.488 .238^ .01BEM YL-1.07+.02X^.567 .321^.083^.01n = 82 pairsp \< .0595Results of the regression analyses for mothers weresimilar to those found for menstrual daughters, revealingthat the predictor variable of self-esteem made anindependent and significant contribution to the variance inbody image scores (23.8%). Therefore, results for mothersrevealed that like menstrual daughters, high self-esteem wasassociated with positive body image. Because self-esteemcontributed to a significant degree to variance in bodyimage scores, the null hypothesis was rejected, indicatingthat scores on self-esteem were found to predict scores onbody image at a significant level.To test Hypothesis II (B), a step-wise regressionanalysis was conducted on the scores obtained by mothers ofmenstrual daughters. Results revealed that the independentvariables together contributed to variance in body imagescores by 32.17%. An F test conducted on the change in Rsquared when gender role identity was added was significantat the p < .01 level, F = (2,79) = 9.29, p < .01. Because itwas determined that self-esteem accounted for 23.8% of thevariance, it was concluded that gender role identity made anindependent and significant contribution to variance in bodyimage scores of 8.3%. Results therefore revealed that highscores on femininity were linked with positive body image.As in the sample of menstrual daughters, self-esteem wasfound to make a greater contribution to variance in body96image scores than did gender role identity. As a result, therelationship between high self-esteem and positive bodyimage was viewed as being stronger than the relationshipbetween positive body image and femininity. Becausesignificant contributions were found to be made by genderidentity in the prediction of body image scores at a p < .05level, the null hypothesis was rejected, indicating thatscores on gender role identity were found to predict bodyimage scores at a significant level.In summary, for Sample I consisting of mothers andmenstrual daughters, scores for both mothers and daughtersrevealed that self-esteem was the strongest predictor ofvariance in body image scores, particularly for daughters.The strength of gender identity in accounting for variationin body image scores was greater for mothers than fordaughters.To test Hypothesis III, a correlation matrix wasproduced in order to determine the relationships betweenspecific mother/menstrual daughter pairs on the predictorand criterion variables. Table 9 presents Pearsoncorrelation coefficients and levels of significance formothers and daughters in Sample 1. The correlation matrixindicated that mothers' scores on body image were positivelyrelated to menstrual daughters' scores on body image (r =97TABLE 9Sample IICorrelation Matrix of BSRI, ROS, & SIQYAfor Mothers and Menstrual Daughters (a) andMothers and Pre-Menstrual Daughters (b)PEARSON CORRELATION COEFFICIENTSa\b^BSRI/M ROS/M SIQYA/M BSRI/D ROS/D^SIQYA/DBSRI/M 1.000 .052 -.109 -.075 -.030^.119P =^. P =^.391 P =^.279 P =^.344 P =^.434 P=.262ROS/M^-.297* 1.000 .379* -.167 330*^.088P =^.003 P =^. P =^.018 P =^.184 P =^.035 P=.317SIQYA/M .130 .488* 1.000 -.293 .077^-.288P =^.122 P =^.000 P =^. P =^.055 P =^.339 P=.108BSRI/D^.098 -.098 -.156 1.000 .202^.364*P =^.190 P =^.190 P =^.080 P =^. P =^.137 P=.022ROS/D^.093 .195* .200* -.016 1.000^.202P =^.202 P =^.039 P =^.036 P =^.441 P =^.^P=.137SIQYA/D^.019 .557* .275* -.209* .577*^1.000P =^.431 P =^.000 P =^.006 P =^.029 P =^.000 P=.a: n = 82 pairsp ‘ .05b: n = 31 pairsp 4 .0598.275; p = .006). Thus it was determined that as mothers'scores rose toward a more positive body image, so did theirdaughters'. Because a significant correlation was revealedbetween body image scores for mothers and menstrualdaughters at a p ‘< .05 level, the null hypothesis wasrejected.The correlation matrix also revealed a positivecorrelation between mothers' and menstrual daughters' scoreson self-esteem (r = .195; p =.039). These findings areconsistent with the results of the multiple regressionanalyses. Also as might be expected from the results of theregression analyses, positive correlations were foundbetween menstrual daughters' scores on body image and self-esteem (r = .577; p = .000) and between mothers' scores onbody image and self-esteem (r = .488; p = .000).A significant negative correlation was found betweenmothers' scores on self-esteem and gender role identity (r =-.297; p = .003). It was therefore determined that anegative relationship existed between the two predictorvariables, so that as scores on self-esteem rose, scores ongender role identity fell. Results of the regressionanalyses for this group may therefore have been tempered bythe finding of multicolinearity. However, the correlationbetween variables was moderate, and did not appear inmothers in Sample II. As well, regression analyses conducted99on scores from the independent variables alone revealed nosignificant effects. It is possible, therefore, that thefinding of multicolinearity was an anomaly. Nevertheless,cautions in the interpretation of the regression analysesfor this group must be taken in the light of this finding.Significant correlations were not found between mothers'scores on gender identity and scores for daughters on thismeasure.To test Hypothesis IV (A), a step-wise multipleregression analysis was conducted on the scores obtainedfrom pre-menstrual daughters on the three instrumentsadministered. Table 6 presents scores for both mothers anddaughters in Sample II, including means and standarddeviations. Results of the regression analysis for daughtersare presented in Table 10. Results indicate that for pre-menstrual daughters, self-esteem did not make a significantcontribution to variance in body image scores. The nullhypothesis was therefore accepted, indicating that scores onself-esteem were not found to predict body image scores forthis group.Regression analyses undertaken to test Hypothesis IV(B)indicated that the independent variable gender role identitymade a significant and independent contribution to theprediction of body image scores (13.3%). Thus, it wasrevealed that for pre-menstrual daughters, femininity was100associated with positive body image. This result isdifferent from that obtained from daughters in Sample I,which identified self-esteem as the strongest predictor ofbody image. Implications of these findings will be discussedin the Discussion section of this paper. Because thecontribution made by gender identity to the prediction ofbody image scores was found to be significant at the p^.05level, the null hypothesis was rejected.To test Hypothesis V (A), a step-wise multipleregression analysis was conducted on scores obtained frommothers of pre-menstrual daughters on the three instrumentsutilized in the study. Results of the regression analysisare presented in Table 11. Results for mothers in Sample IIwere found to be similar to those obtained from mothers inSample I: the regression analysis revealed that self-esteemaccounted for 14.4% of variance in body image scores.Because this result was found to be significant at a p^.05level, the null hypothesis was rejected.To test Hypothesis V (B), regression analyses wereconducted on scores obtained by mothers of pre-menstrualdaughters. Results indicated that gender role identity didnot make a significant contribution to the variance in bodyimage scores for these women. The null hypothesis wastherefore accepted.101TABLE 10Sample IIDaughtersStep-wise Multiple Regressions on Daughterson Independent VariablesIndependent^RegressionVariable^Equation^R^R Square PBSRI^YI=1.97+.04x^.364 .133^.044n = 31 pairsP^.05TABLE 11Sample IIStep-wise Multiple Regressions on Mothers onIndependent VariablesIndependent RegressionVariable^Equation R^R Square PROS^Y62.67+.58x^.379 .144^.035n = 31 pairsp < .05102To test Hypothesis VI, a correlation matrix wasproduced in order to determine the relationships betweenspecific mother/pre-menstrual daughter pairs on thepredictor and criterion variables. Table 9 presents Pearsoncorrelation coefficients and levels of significance formothers and pre-menstrual daughters on the three variables.A significant correlation was not found for body image inthese mothers and their daughters. A significant positivecorrelation was found, however, for self-esteem betweenthese mothers and their daughters (r = .330; p = .035).Thus, results indicated that as mothers' scores on self-esteem rose, so did their daughters'. Consistent with theresults of the multiple regression analysis, the correlationmatrix revealed a significant positive correlation formothers between self-esteem and body image (r = .379), p =.018. This result indicated that for mothers of pre-menstrual girls, a positive relationship existed betweenscores on body image and self-esteem, so that increases inbody image scores were related to increases in scores onself-esteem. Also consistent with the results of themultiple regression analysis was the finding of asignificant positive relationship for daughters betweengender identity and body image (r = .364; p = .022), withincreases in body image scores related to increasing scoreson femininity. Since a significant correlation was not found103between mothers and daughters on the dependent variable,body image, the null hypothesis was accepted.Following the above procedures, a series of t-testswere undertaken as post hoc exploratory measures. The firstt-test was conducted in order to determine whether bodyimage scores for all mother/daughter pairs were related toparticipants' categorical ratings (ie, feminine, masculine,androgynous, and undifferentiated) on the Bem Sex-RoleInventory (Bem, 1981). Significant results were not found. Asecond t-test was conducted in an effort to determinewhether mothers' employment status in the first five yearsof their daughters' lives was related to daughters' bodyimage scores. Again, significant results were not found. At-test was also undertaken which compared the scores of asub-sample of 14 menstrual and 12 pre-menstrual girls aged12 years. This test did not reveal significant results.Finally, no significant results were found on a t-test whichexamined the relationship between various categories ofreligious orientation and daughters' scores on body image.Standardized scatterplot diagrams were produced for alldata obtained in the study in order to ensure normaldistribution of scores. All diagrams revealed normaldistributions (see Appendix H).104CHAPTER FIVEDiscussionIn the present chapter, a restatement of the purpose ofthe study is presented, followed by a discussion of theparticipant sample and a summary and discussion of theresults obtained from testing the hypotheses. A discussionof implications for future research and recommendations forcounselling will conclude the chapter.Restatement of the Purpose The primary purpose of the study was to explore therelationship between self-esteem, gender role identity, andbody image for mothers and their adolescent daughters duringthe pubertal period of development. It was hypothesized thatthe relationships between variables evident in mothers wouldbe comparable to the relationships evident in daughters.Initially, it was decided that the cultural, social,biological, and developmental significance of menarcherequired that the adolescent girls selected for inclusion inthe study had experienced the onset of menarche within 36months of the time of data collection. However, completedtest packages returned allowed the researcher to include asecond sample of mother/daughter pairs in which thedaughters had not yet experienced menstruation. This allowedfurther comparisons to be made between pre-menstrual and105menstrual girls. As a result, the primary purpose of thestudy was expanded to include an exploration of thedifferences between the two groups on the relationshipsbetween variables.Discussion of the Sample The sample reflected a well-educated, middle to upper(dual) income population, who were not highly committed toorganized religion. The majority lived in traditional familysettings, with mothers working in traditional occupations.Despite the lack of random sampling, participants weresomewhat diverse in terms of religious affiliation andethnicity.Because the sample consisted of predominantly whiteupper to middle class women and girls, it is possible thattheir responses reflected the cultural attitudes and valuesof this group evident in larger society. Comparisons made onthe basis of culture could not be made in this sample group.As well, the high level of education evident in the sampleof mothers may indicate a sample of women who are aware ofthe issues surrounding the variables under investigation.Their responses may reflect a heightened awareness to theseissues as compared to the population in general. Finally,the financial stability, and the traditional family settingscharacteristic of this sample may have affected responses in106the direction of more positive ratings. It is possible thatthe participants' environment had an influence on the typeof response. That is, the environmental influences, and thetypes of experiences participants had as a result of theirenvironment may have affected their perceptions ofthemselves, and therefore, their responses. However, thenature of the impact was not investigated in the presentstudy.Overall, scores for both groups of mothers anddaughters on self-esteem were quite positive, indicating afairly high level of self-esteem for the two samples ingeneral (Tables 5 & 9). As well, scores on body image weregenerally high (Tables 5 & 9). Gender role identity scoreswere consistent with the scores obtained for comparable normgroups (Bem, 1981). Scores therefore reflected a sample ofparticipants who had generally positive perceptions abouttheir bodies, had generally high self-esteem, and who ratedtheir gender role identity in a similar manner to theircounterparts in norm groups.Demographic differences were noted between the samplegroups. For example, significant differences were foundbetween daughters in Samples I and II (Table 4). Daughtersin Sample I, the menstrual girls, were found to besignificantly older, taller, and heavier than daughters inSample II. As well, daughters in Sample I were found to be107in a higher school grade. These variables areinterconnected, and reflect developmental differencesbetween daughters in the two groups. For example, menstrualgirls are more likely to be older, taller, and heavier thantheir premenstrual counterparts. Although these differenceswere found to be significant, they were relatively small(see Table 4). As well, results of a t-test which comparedscores of 12 year old girls in both samples revealed nosignificant differences. This result indicates that therelationships evident between variables may be related tomenstrual status rather than age. Thus, although differencesin age, height and weight may account for some of thevariations in responses between the two groups, it is alsopossible that the onset of menstruation played a role in thedifferences evident in responses between daughters in thetwo sample groups.Significant differences were also found between mothersin Sample I and Sample II. T-tests revealed that mothers inSample II were older than mothers in Sample I, and thatmothers in Sample II had been employed in the first fiveyears of their daughters' lives at a higher rate thanmothers in Sample II (see Table 3). Although the differencein age between mothers is small, it may have affectedresponses, as it indicated a greater age difference betweenthe younger, pre-menstrual girls and their older mothers. As108well, the greater amount of time spent by mothers of pre-menstrual girls in their first years of life may haveresulted in a closer, more intimate mother-daughterrelationship for these participants. However, this study didnot investigate the impact of these differences in detail.In summary, the two samples were similar in terms ofclass, family setting, income, and religious orientation.Small but significant differences were found between thesample groups, which may have affected the outcome of thestudy. The present study did not explore the implications ofthese differences in detail.Summary and discussion of results Sample I Testing of the first hypothesis indicated that formenstrual daughters, self-esteem accounted for 33.3% ofvariation in body image scores, followed by gender roleidentity which accounted for an additional 4.0% of thevariation in body image scores. Results for mothers ofmenstrual girls were similar to those for daughters: self-esteem was found to predict body image (23.8%), followed bygender role identity, which added to variance in body imagescores by an additional 8.3%.The relationship between self-esteem and body imageevident in the present study is consistent with results ofprevious research on both women and adolescent girls109( Jackson et al., 1988; Jones & Mussen, 1958; Lerner et al.,1976, Noles et al., 1985; Rosen & Ross, 1968; Secord &Jourard, 1953). Like the research cited above, in this studyhigher self-esteem scores were found to correspond to morepositive body image scores. The strength of self-esteem inpredicting body image was greater for daughters (33.3%) thanfor mothers (23.8%). By demonstrating a significantrelationship between these variables, this result appears toadd support to research which links body image to self-esteem, particularly in the context of adolescentdevelopment (Attie & Brooks-Gunn, 1989; Fabian & Thompson,1989; Koff et al., 1978; Petersen, 1988; Rierdan & Koff,1980; Rosenbaum, 1979).Testing of the first and second hypotheses alsoidentified gender role identity as a significant predictorof body image. However, for daughters, it was determinedthat negative body image was associated with greaterfemininity (4.0%), while for mothers, positive body imagewas related to higher scores for femininity (8.3%). Previousresearch has indicated that for women, perception offemininity is related to negative body image ( Jackson etal., 1987, 1988; Kimlicka et al., 1983; Winstead & Cash,1984). Thus, for menstrual girls in the present study,results are consistent with evidence presented in earlierresearch. For mothers of menstrual girls, however, the110relationship existed in the opposite direction: higherscores on femininity were associated with more positive bodyimage scores. For this group, higher femininity scores wererelated to both higher self-esteem and more positive bodyimage scores. This result may indicate developmentaldifferences between mothers and daughters. That is,adolescent girls may be demonstrating that they may beexperiencing a process of identity formation, in which thephysical characteristics of femininity, such as breastdevelopment and onset of menses are not yet integrated withthe psychological aspects of identity development. Formothers, however, the integration of physical andpsychological components of identity may have occurred,leading to a more congruent relationship between femininity,self-esteem, and body image. The relationship of higherfemininity and negative body image evident in menstrualgirls may also be associated with the negative attitudetoward menstruation expressed by the adolescent participantsin the present study (to be discussed in the followingsection). For these girls, then, negative perceptions of thephysical changes that accompany the onset of menses may beassociated with femininity, and therefore, with a morenegative body image.The third hypothesis stated that for mother/menstrualdaughter pairs, mothers' body image scores would not predict111daughters' body image scores. A significant positivecorrelation was found between mothers and daughters on thedependent variable (r = .275, p = .006). As a result, thenull hypothesis was rejected. A positive correlation wasalso found between mothers' self-esteem and that ofdaughters (r = .195, p = .039). However, a significantcorrelation was not found for mothers and daughters on thevariable of gender identity.This finding lends support to earlier evidenceidentifying the mother-daughter relationship as asignificant predictor of the adolescent girl's self-image,particularly at puberty ( Holmbeck & Hill, 1986: Leaper etal., 1989; LeCroy, 1988; Offer et al., 1982). Specifically,the present study suggests that there may be an associationbetween the body image of an adolescent girl and that of hermother.The finding of a positive correlation between mothersand their menstrual daughters on body image and self-esteemis consistent with the theoretical framework offered byobject relations theorists, who emphasize the centralimportance of the mother-daughter relationship in adolescentidentity formation (Bassof, 1988; La Sorsa & Fodor, 1990;Miller, 1976; Rubin, 1983). In the present study, thepositive correlation between mothers and daughters on bodyimage and self-esteem scores might be indicative of the112presence of an interactional process between mother anddaughter. That is, results of the present study indicatedthat higher scores on body image and self-esteem for motherswere associated with higher scores on these variables fordaughters. By identifying a positive correlation betweenbody image and self-esteem scores for mothers and daughters,the present study adds empirical support to the theoreticalposition that the mother-daughter relationship is a criticalcomponent in adolescent development, including thedevelopment of body image, and thereby emphasizes thereciprocal influence within the mother-daughter relationship(Bassof, 1988; La Sorsa & Fodor, 1990; Rubin, 1983). Becausethe present study did not investigate alternate factors,such as the influence of the father or peer relationships,it may be that these or other factors may have contributedto the finding of a correlation between mothers' anddaughters' scores on body image. Further study, whichidentified and explored these factors would be required inorder to gain a better understanding of the extent to whichthe mother-daughter relationship may influence thedevelopment of the adolescent girl.In the present study, although high self-esteem scoreswere linked to high femininity scores for both mothers andmenstrual daughters, a significant correlation was not foundbetween mothers and daughters on gender role identity.113Differences between mothers and daughters were evident inthe strength of the contribution made by gender roleidentity to variance in body image scores (Tables 10 & 11).Thus for mothers, gender role identity contributed tovariance in body image scores by 8.3%, while for menstrualdaughters, gender role identity contributed only 4.0%. Formothers, then, gender role identity was more closelyassociated with self-esteem and body image than formenstrual daughters. This result is consistent with theresults showing no significant correlation between genderrole identity scores for mothers and daughters.This is perhaps explained by the arguments of La Sorsa& Fodor (1990), who posit that recent societal changes inroles and expectations for women may act to complicate themother-daughter relationship further:adolescent daughters of these [contemporary] mothersare the first generation that from birth was exposed tothe expanded options for women; that is , the firstgeneration of women who could aspire to being more thanwives and mothers or entering traditional women'soccupations. (p.594)Unlike earlier generations, the contemporaryadolescent's identity crisis may not be resolved through themodeling of her mother's traditional role (Erickson, 1968).Instead, models include adult women (including mothers) who114demonstrate a much wider variety of options. La Sorsa &Fodor (1990) argue that the increased options for women,often modelled by a mother who is an active participant inthe work force, may lead to confusion and a prolonged periodof identity formation for the contemporary adolescent girl.Results of the present study may therefore indicategreater confusion and a lack of integration in gender roleidentity formation by daughters. It is also possible thatthe weaker strength of gender role identity in predictingbody image for daughters may indicate that daughters, unlikemothers, are still in the process of developing gender roleidentity. For daughters, then, gender role identity may bevague and poorly defined. As well, it may be that fordaughters, the associations between gender role identity andactual behavior may not be as rigid and prescribed as formothers. Thus, because they belong to a different generationof women, for mothers the issue of gender role identity maybe more salient and more closely associated with overallidentity, than for daughters.Sample II The fourth hypothesis (A & B) tested stated that forpre-menstrual girls, self-esteem and gender role identitywould not predict body image as measured by the instrumentsindicated above. Results identified gender role identity asthe only significant predictor of body image (13.3%). As a115result, Hypothesis IV(A) was rejected and Hypothesis IV(B)was accepted.These results are consistent with the findings ofresearch comparing menstrual and premenstrual girls (Koff etal., 1978; Rierdan & Koff, 1980). In this body of research,it was reported that pre-menstrual girls were more concernedwith the physical characteristics of puberty, such as breastdevelopment, than menstrual girls. Thus for pre-menstrualgirls in the present study, it may be that feminine genderrole identity was perceived or defined in terms of thephysical changes of puberty. This concern with physicalchanges may reflect the association between gender roleidentity and body image for pre-menstrual girls in thepresent study.The difference in results between pre-menarchal andmenarchal girls provides additional support to researchwhich identifies the onset of menarche as a pivotal eventaround which the adolescent girl's body image and sexualidentification come to be re-organized ( Koff et al., 1978;Rierdan & Koff, 1980, 1985). The nature of the differencesbetween daughters in Samples I and II, and the implicationsof these differences, will be discussed at the conclusion ofthis section.Hypothesis V (A & B) stated that for mothers of pre-menstrual daughters, self-esteem, as measured by the116Rosenberg Self-Esteem Scale, and gender role identity, asmeasured by the Bem Sex-Role Inventory, would not predictbody image. Testing of the hypothesis revealed that self-esteem contributed to variance in body image scores to asignificant degree (14.4%), while gender role identity didnot make a significant contribution. Null Hypothesis V(A)was therefore accepted and null hypothesis V(B) rejected.For mothers of pre-menstrual girls, gender roleidentity was not found to make a significant contribution tovariance in body image scores. This could be attributed tothe small size of the sample ( n = 31). As well, this resultmay have emerged because unlike the daughters in Sample I,the daughters of this sample were pre-menstrual andsignificantly younger than menstrual daughters (Table 4).Results of the present study may therefore indicate thatgender role identity may be a more salient issue for mothersof menstrual daughters. In contrast, for mothers of younger,pre-menstrual daughters, gender role identity may not yethave re-emerged as a relevant or meaningful issue. Althoughresearch has not investigated this area in detail, there isevidence in the literature that mothers and daughtersinteract differently according to the daughter's menstrualstatus (Hill et al., 1985). This research suggests that theperiod following the onset of menstruation is characterizedby stress and often conflict between mother and daughter. It117is arguable, therefore, that the onset of menstruation maytrigger the emergence of developmental issues such as genderrole identity for both mothers and daughters. Results of thepresent study may indicate that mothers of pre-menstrualgirls may not yet have been confronted by this issue.The testing of Hypothesis VI revealed that formother/pre-menstrual daughter pairs, a significantcorrelation was not found between mothers and daughters onthe dependent variable, body image. The null hypothesis wastherefore accepted.The lack of a significant correlation between bodyimage scores for mothers and pre-menstrual daughters may bedue to the small sample size (n = 31 pairs). However, theseresults may also be due to the developmental status of thedaughters. Because the sample of pre-menstrual girls wassignificantly younger and smaller in physical size than thesample of menstrual girls (Table 4), it is arguable thatthis population tended to be less developed or physicallymature. As a result, these mothers and daughters may not yethave been confronted by the developmental challengestriggered by pubertal changes described by developmentaltheorists and researchers ( Attie & Brooks-Gunn, 1989; LaSorsa & Fodor, 1990; Petersen, 1988; Rierdan & Koff, 1980).The salience of body image, and the manner in which it is118assessed, may therefore be quite different for pre-menstrualdaughters than for their mothers.A significant positive correlation between self-esteemscores for mothers and daughters was revealed (r = .330, p =.035). However, a significant correlation between mothersand daughters on gender role identity was not found. It ispossible that this result might be due to the effect of thedaughters' pubertal status on the mother-daughterrelationship. For pre-menstrual daughters, gender roleidentity may be in the process of development, and may bedefined in terms of physical characteristics. It may be thatthe onset of menarche in daughters triggers a process ofreassessment of gender role identity for mothers. Formothers of pre-menstrual girls, perhaps this process has notyet begun. Results of the present study may reflect thatmothers and pre-menstrual daughters are in an earlier phaseof development, and as a result their perceptions ofdevelopmental issues such as gender role identity may bedifferent.For mothers of both mensrtrual and pre-menstrual girls,results of the present study indicated that higherfemininity scores were associated with higher self-esteemand more positive body image. This is in directcontradiction to results of the majority of research in thisarea (Jackson et al., 1988; Whitely, 1983, 1988). The119results of this study, however, may reflect the phenomenonreported by Jackson et al. (1988). Although the authorsfound that in their investigation, women with high scores onfemininity also demonstrated negative body image, theyreported that there were no differences among women in theimportance they ascribed to physical appearance. Theresearchers discussed this finding as evidence of theacceptance among contemporary women of cultural standards onthe importance of appearance for women. In the present studyas well, scores may reflect participants' acceptance ofthese cultural standards, with the result that high scoreson both self-esteem and femininity were associated withpositive body image.Examination of the data reveal notable differencesbetween Sample I and Sample II in the results of theregression analyses. For example, for daughters in Sample I,self-esteem emerged as the strongest predictor of body image( 33.3%) followed by gender identity (4.0%), while forSample II daughters, gender role identity proved to be theonly significant predictor of body image (14.3%). As well,results of t-tests which revealed no significant differencesbetween the scores of a sub-sample of same-aged menstrualand pre-menstrual girls indicated that differences may berelated to menstrual status rather than age. These resultspoint to a noteworthy difference between pre-menstrual and120menstrual girls in the factors which are related to bodyimage, and support increasing evidence presented incontemporary literature that the onset of menarche is acritical experience for the adolescent girl (Attie & Brooks-Gunn, 1989; Fabian & Thompson, 1989; Koff et al., 1978;Petersen, 1988; Rierdan & Koff, 1980; Rosenbaum, 1979).Research on menstruation identifies body image and sexualidentification as the variables which represent the greatestdifferences between pre-menstrual and menstrual girls. Forexample, the present study supports evidence presented byKoff et al. (1978) and Rierdan & Koff (1980) that pre-menarchal girls were more preoccupied with changes in bodyparts, and by extension sexual identification, than theirmenarchal counterparts who exhibited a more integrated bodyimage.Researchers of female adolescent development suggestthat menarchal girls are less pre-occupied with issues ofphysical maturation and sexual identification than pre-menstrual girls; as a result, their body image may be morecoherent and better integrated (Koff et al., 1978; Rierdan &Koff, 1980). Thus, for menstrual girls gender role identitymay be associated with the behavioral and psychologicalaspects of development, rather than with physicalmaturation. The emergence of positive self-esteem as thestrongest predictor of body image in menarchal girls in the121present study suggests that this population may haveexperienced the physical changes of puberty. The salientdevelopmental tasks may therefore entail integration ofthese changes into overall identity. Thus, menarchal girlsmay begin to make associations between body image and self-esteem. The present study indicated that for menarchalgirls, the relationship between body image and self-esteemis similar to the relationship between these variablesevident in adult women, with higher self-esteem beingrelated to more positive body image (Jackson et al., 1988;Jones & Mussen, 1958; Lerner et al., 1976; McCaulay et al.,1978; Noles et al., 1975). This result suggests thatmenarchal girls may be moving toward a more mature phase ofdevelopment than their pre-menarchal counterparts.The differences between pre-menstrual and menstrualgirls evident in the present study emphasize the importanceof the onset of menarche in the developmental process ofadolescence. The differences evident between the groupssuggest that for adolescent girls, the onset of menarche maybe associated with shifts in self-esteem, gender roleidentity, and body image, and may affect the relationshipsbetween these variables. In addition, the evidence suggeststhat the onset of menarche may have implications for thedevelopment of the mother as well as the daughter. However,the present study did not explore alternate factors which122may have contributed to differences between the groups, suchas whether the girl perceived her development to be early,on time or late compared to her peers. Clearly, furtherstudy is required if this process is to be understood ingreater depth.Comments by Participants on Menstruation The participating daughters were not asked to describethe meaning menarche had for them; however, many offeredtheir comments, which were overwhelmingly negative. Forexample, in reply to the question of when they hadexperienced their first menstruation, responses included thefollowing:Not yet! And glad of it!It's good to have the first at 13. It's kind of late ,I guess, but the later the better, more educated aboutit and have less no. of times.Not a great experience.I was really scared and freaked out, now I'm fine.I was so scared.I really wish I didn't have it. It really gets in theway with my social life. I would have liked it to comein a few years or so.I hate it! It really sucks! Why don't guys get it.Too early.Hated it!123These responses seem to reflect western society'sattitudes toward menstruation, which to date tend to becharacterized by shame and embarrassment (Delaney et al.,1988). The present study was not designed to explore themeanings and implications of menarche for mothers anddaughters, nor did the researcher attempt to discover theeffect of the onset of menarche on the mother-daughterrelationship. Comments made by participants indicate thatthese issues are relevant and often problematic forcontemporary adolescent girls.Reactions by Participants to Study The present study was designed as an exploratoryinvestigation of relationships between the variables ofself-esteem, gender role identity and body image inadolescent girls and their mothers. The design reflected thepurpose of the study: data were collected and analyzed in aneffort to explore relationships between variables.Instruments were chosen on the basis of their reliabilityand validity, as well as their ease of distribution andcompletion. As such, the design and instrumentation weredetermined to meet the requirements of the investigation toa satisfactory degree.Some respondents expressed a certain amount offrustration, both with the design of the study and theinstruments utilized. For example, the researcher received a124number of telephone requests for greater involvement in thestudy; participants stated that the instruments limitedtheir ability to express themselves. Because in the majorityof cases daughters were contacted first, some parentsexpressed a desire for direct contact with the researcherprior to participation, by telephone and or in writing.Reactions among the majority of participants to thestudy was enthusiastic, and often interested and curious.Some participants contacted the researcher by telephone torequest details on the purpose and scope of the study and tooffer their participation in future research in this area.Most requested that the results of the study be madeavailable to them. It was concluded, therefore, that thetopic of body image, particularly in the context of themother-daughter relationship was of considerablesignificance to those contacted.The level of enthusiasm expressed by many participants,and their requests for the results in the study wasinterpreted as an indication that participants felt that thesubject matter was relevant and had significance in theirlives. Daughters were most articulate in their responses tothe study. For example:I hope your research is successful because it soundslike a good topic.125I think that this is an important study. I, myself haveoften been unsatisfied with myself physically but sofar I haven't been so desperate as to throw up oranything. I know my friends haven't always been solucky.Finally someone does this! Good idea.This was a fun quiz. At times I found it hard toanswer.I think that studies like these are very important inhelping me and others realize that each person in theworld is different, mentally, emotionally, andphysically. Thanks.I think that this test is good at making people seewhat they don't like about themselves.This study interests me greatly. Please continue tosend me the questionnaires. Thanxs!Implications for Further ResearchBecause this study was exploratory in nature, it wasuseful in identifying critical issues in the area of bodyimage development. A comprehensive study of this area wouldrequire random sampling and more rigorous design, includingboth quantitative and qualitative methods, so that theissues of body image, self-esteem, and gender role identitymight be addressed in depth. Recommendations for futureresearch are as follows:126(1) Based on the finding that for adolescent girls, self-esteem and gender role identity are related to body image,more research which investigates the relationships betweenthese variables in further depth is required. Thus researchis required which allows the researcher to examine theextent to which these variables affect the development ofbody image. Traditionally, body image research has involvedthe use of quantitative measures which attempt to measureparticipants' degree of distortion in body size. A designwhich explores the affective component of body image, andits relation to other variables such as self-esteem wouldaddress the dearth of research and understanding in thisarea. The development of a valid and reliable measure whichtaps the affective component of body image for both femaleadolescents and adult women is therefore recommended.Alternatively, a qualitative method which allowsparticipants to express their perceptions in detail wouldprovide investigators with valuable data on the relevance ofbody image, and how it is related to the variablesidentified in the present study.(2) Based on the differences between pre-menstrual andmenstrual girls identified by the present study, research onthe impact of the onset of menarche on body image, and onoverall identity development is required for a greaterunderstanding of female adolescent development. It is127recommended that this research is undertaken in the contextof the mother-daughter relationship. A longitudinal design,which allows the researcher to follow adolescent girls andtheir mothers through the developmental process surroundingthe onset of menarche is therefore recommended.(3) Based on the finding of a significant relationshipbetween mothers' and menstrual daughters' scores on bodyimage and self-esteem, further research is recommended whichinvestigates the interactional nature of the development ofthe midlife mother-adolescent daughter. A structuredinterview approach, which allows the researcher toinvestigate the individual and interactional processes ofadolescent daughter-midlife mother pairs would provideinformation on body image in the context of the relationshipbetween mothers and daughters. Alternatively, aquantitative, longitudinal design is recommended so that therelationships between the previously identified variablesmight be examined as mid-life mothers and adolescentdaughters experience the developmental transitions of thisstage.Because body image is a vital component in identityformation, research in this field has far-reachingimplications for many fields, including eating disorders,sexuality, relationships, and education. As such, it mayserve to aid professionals and clinicians in many areas, not128only in terms of the treatment of disorders, but inprevention of potential difficulties, particularly thoseencountered by mothers and daughters as they struggle toaddress and resolve developmental tasks and transitions. Anunderstanding of the relationships between variables wouldallow clinicians to view body image in the context of therelationships that exist with self-esteem and gender roleidentity. In this way, various aspects of development suchas body image or self-esteem may be addressed individually,or in terms of their relationships within a developmentalcontext.Recommendations for CounsellingThe present study has implications for familycounsellors and counsellors of adolescent girls. Based onthe evidence demonstrating relationships between mothers anddaughters on the variables of body image and self-esteem,counsellors of adolescent girls may better address theirclients' concerns in the context of the mother-daughterrelationship. As such, the counsellor may choose to includethe mother in counselling sessions. For example, thisapproach may be of use in situations where both mother anddaughter are experiencing difficulty in dealing with thedaughter's sexuality. Issues of body image and self-esteemmay be salient for both mother and daughter, as well asissues concerning the daughter's increasing independence and129autonomy. Inclusion of the mother in counselling sessions,along with an awareness of the implications of these issuesfor both mother and daughter, may allow the counsellor tofacilitate the developmental processes of both clients byaddressing the developmental issues salient to mother anddaughter. In this way, the needs of both mother and daughtermight be met, and the relationship itself enhanced.Counsellors who encounter adolescent clientsexperiencing difficulties in facing developmental tasks suchas adjustment to pubertal changes may integrate informationprovided by this study into their interventions. Forexample, in addressing the issue of negative body image, thecounsellor may investigate the client's menarchal status,self-esteem, and gender role identity. Sensitivity to issuesof body image may allow the counsellor to detect potentialdifficulties in various areas, including the area of eatingdisorders, and to intervene accordingly. As well, thecounsellor may use this information to anticipatedifficulties in younger clients, and therefore be able totake preventative measures. For example, a counsellor mayrecognize difficulties in body image or gender role identityin a pre-menstrual girl, and may use this information toaddress the issues surrounding the onset of menstruation inorder to offset potential difficulties in the future.130The negative comments concerning menstruation offeredby the adolescent participants in this study indicate thatthe onset of menstruation may be a source of shame and fearfor many girls. This information points to a need forcounselling interventions to address these attitudes.Targets for interventions may include pre-menstrual girlsand girls who have recently experienced the onset ofmenarche so that they may express their attitudes and fears,and be offered support, validation, and alternative pointsof view. As well, adolescent boys may benefit by counsellingor education on the topic of menstruation, in order to de-mystify the experience of menstruation, and to allow boys tofeel included and therefore less fearful of the process andits implications. Finally, family counselling may aid familymembers, particularly mothers, to express their attitudes,and perhaps may encourage families to support theiradolescent girls in aquiring a positive and healthy attitudetoward menstruation.Finally, it is recommended that counsellors andeducators implement information provided by previousresearch and the present study in order to prevent potentialdifficulties among adolescents in the areas of body imageand self-esteem. An understanding of the various componentswhich may affect adolescent development and how thesecomponents are related, may be used to educate young131adolescents as they enter this stage. For example,information on the physical and psychological effects of theonset of menarche may aid adolescent girls to address issuesof body image and gender role identity by providing insightinto the normal developmental changes and theirimplications. Adolescent girls may then be able toanticipate developmental tasks with a greater understandingof how these changes may affect them physically andpsychologically. In this way, the counsellor or educator mayuse available information to prevent potential difficultiesfrom arising. As well, this information may be presented toolder adolescents in the classroom or in counsellingsettings, in order to provide insight into the nature of thedevelopmental processes they are currently experiencing, andto open an avenue for discussion. In this way, adolescentsand their families may be better equipped to address issuesrelated to the difficult and sometimes problematicdevelopmental transitions of adolescence.Conclusion The results of the present study have raised severalissues. First, they support evidence presented throughprevious research that self-esteem and gender role identityare related to body image for adult women and adolescentgirls. That is, results obtained in the present studyindicate that for mothers and menstrual daughters, higher132self-esteem and higher ratings on femininity were related tomore positive body image. Second, the relationships evidentbetween variables were different for menstrual and pre-menstrual girls. This study therefore identified differencesin the relationships between variables based on menarchalstatus. For pre-menstrual girls, gender role identity wasfound to be the only variable associated with body image.Like menstrual girls, higher femininity was found to beassociated with more positive body image for this sample.Finally, evidence was presented that revealed relationshipsbetween mothers' and daughters' scores on the abovevariables. In particular, a positive correlation was foundfor body image and self-esteem scores between mothers andmenstrual daughters. These results support findings fromprevious research that the mother-daughter relationship maybe a critical component of the daughter's development,particularly at adolescence.On the basis of the findings of this study and ofprevious studies, it can be argued that self-esteem, genderrole identity, and menarchal status are important componentsin the development of body image at adolescence.Furthermore, the dynamic interactional nature of the mother-daughter relationship seems to act as a vital context inwhich overall identity, including body image develops. Thenature of the relationships between these factors may.133therefore affect the developmental processes of both motherand daughter.134REFERENCESAttie, I., & Brooks-Gunn, J. (1989). Development of eatingproblems in adolescent girls: A longitudinal study.Developmental Psychology, 25, 70-79.Bassof, E. (1988). Mothering adolescent daughters: Lovingand letting go. New York: American Library.Bem, S. L. (1979). Theory and measurement of androgyny: Areply to the Pedhazur-Tetenbaum and Locksley-Coltencritiques. Journal of Personality and Social Psychology, 37, 1047-1054.Bem, S. L. (1981). Gender schema theory: A cognitive accountof sex typing. Psychological Review, 88, 354-364.Blyth, D. A., Simmons, R. G., & Zakin, D. F. (1985).Satisfaction with body image for early adolescentfemales: The impact of pubertal timing within differentschool environments. Journal of Youth and Adolescence,14 , 207-223.Boskind-Lodahl, M. (1976). Cinderella's stepsisters: Afeminist perspective on anorexia nervosa and bulimia.Signs: Journal of Women in Culture and Society, 2, 342-356.Brown, J. A., Cross, H. J., & Nelson, J. M. (1990). Sex-roleidentity and sex-role ideology in college women withbulimic behavior. International Journal of EatingDisorders, 5, 571-575.135Brooks-Gunn, J., Rock, D., & Warren, M.P. (1989).Comparability of constructs across the adolescentyears. Developmental Psychology, 25, 51-60.Bruch, H. (1978) The golden cage. Cambridge: HarvardUniversity Press.Cantelton, L. J., Leichner. P. P., Harper, D. W. (1986)Sex-role conflict in women with eating disorders.International Journal of Eating Disorders, 5, 317-323.Cash, T. F. (1981). Physical attractiveness: An annotatedbibliography of theory and research in the behavioralsciences. JSAS Catalogue of Selected Documents in Psychology, 11, ( MS.2370).Clifford, E., (1971). Body satisfaction in adolescence.Journal of Perceptual and Motor Skills, 33, 119-125.Coopersmith, S. (1967). The antecedents of self-esteem. SanFrancisco: Freeman.Crockett, L. J., & Petersen, A. C. (1987). Pubertal statusand psychosocial development: Findings from the EarlyDevelopment Study. In R. M. Lerner & T. T. Foch (Eds.),Biological-psychosocial interactions in early adolescence: A life-span perspective (pp. 173-188).Hillsdale, NJ: Erinbaum.Davies, E. & Furnham, A. (1986). Body satisfaction inadolescent girls. British Journal of Medical Psychology, 59, 279-287.136Delaney, J., Lupton, M.J., & Toth, E. (1988). The curse: Acultural history of menstruation. University ofIllinois Press: Chicago.Erikson, E. (1950). Childhood and society. New York: W.W.Norton.Etough, C., & Weber, S. (1982). Perceptions of sex-roles inyoung and middle-aged men and women. Perceptual andMotor Skills, 55, 559-562.Fabian, L. J., & Thompson, J. K. (1989). Body image andeating disturbance in young females.  International Journal of Eating Disorders, 8, 63-74.Falbo, T. (1977). Relationships between sex, sex role, andsocial influence. Psychology of Women Quarterly, 2, 62-72.Feldman, S. S., Biringen, Z. C., & Nash, S. C. (1981).Fluctuations of sex-related attributions as a functionof stage of family life. Developmental Psychology, 17,24-35.Fisher, S., & Cleveland, S. (1968). Body image andpersonality. New York: Dover Publications.Fleck, J. R., Fuller, C. C., Malin, S. Z., Miller, D. A., &Acheson, K. R. (1980). Father psychological absence andheterosexual behavior, personal adjustment, and sex-typing in adolescent girls. Adolescence, 15, 847-860.137Frable, D. (1989). Sex typing and gender ideology: Twofacets of the individual's ideology that gotogether. Journal of Personality and Social Psychology, 56, 95-108.Franzoi, S.L., & Herzog, M. E. (1987). Judging physicalattractiveness: What body aspects do we use?Personality and Social Psychology Bulletin, 9, 13-19.Fazier, A., & Lizonbee, L. K. (1960). Adolescent concernswith physique. The School Review, 397-406.Garner, D. M., Garfinkel, P. E., Schwartz, D. E., &Thompson, M. (1980). Cultural expecations of thinnessin women. Psychological Reports, 47, 843-491.Gecas, V. (1972). Parental behavior and contextualvariations in adolescent self-esteem. Sociometry, 35,332-345.Gilligan, C. (1982). In a different voice. Cambridge:Harvard University press.Gold, M., & Yanof, D. S. (1985). Mothers, daughters, andgirlfriends. Journal of Personality and Social Psychology, 49, 654-659.Golombeck, H., Marton, P., Stein, B., & Korenblum, M.(1987). Personality functioning status during early andmiddle adolescence. Adolescent Psychiatry, 14, 365-377.138Gove, W. R., & Herb, T. R. (1974). Stress and mental illnessamong the young: A comparison of the sexes. Social Forces, 53, 256-265.Hawkins, R. C., Turrel, S., & Jackson, C. J. (1983).Desireable and undesireable masculine and femininetraits in relation to students' dieting tendencies andbody image satisfation. Sex Roles, 9, 705-719.Hill, J. P., Holmbeck, G. N., Marlow, L., Green, T. M., &Lynch, M. E. (1985). Menarcheal staus and parent-childrelations in families of seventh-grade girls. Journal of Youth and Adolescence, 14, 301-315.Holmbeck, G. N., & Bale, P. (1988). Relations betweeninstrumental and expressive personality traits andbehaviors: A test of Spence and Helmreich's theory.Journal of Research in Personality, 22, 37-59.Holmbeck, G. N., & Hill, J. P. (1986). A path-analyticapproach to the relations between parental traits andacceptance and adolescent adjustment. Sex Roles, 14,315-334.Isberg, R. S., Hauser, S. T., Jacobson, A. M., Powers, S.I., Noam, G., Weis-Perry, B., & Follansbee, D. (1989).Parental contexts of adolescent self-esteem: Adevelopmental perspective. Journal of Youth andAdolescence, 18, 1-23.139Jackson, L. A., Ialongo, N., & Stollak, G. (1986). Parentalcorrelates of gender role: The relations betweenparents' masculinity, femininity, and child-rearingbehaviors and their children's gender roles. Journal of Social and Clinical Psychology, 4, 204-224.Jackson, L. A., Sullivan, L. A., & Hymes, J. (1987). Gender,gender role, and physical appearance. Journal of Psychology, 121, 151-156.Jackson, L. A., Sullivan, L. A., & Rostker, R. (1988).Gender, gender role, and body image. Sex Roles, 19,429-443.Jones, M. C., & Mussen, P. H. (1958). Self conceptions,motivations, and interpersonal attitudes of early- andlate-maturing girls. Child Development, 29, 491-501.Kamptner, L. N. (1988). Identity development in earlyadolescence: Causal modeling of social and familialinfluences. Journal of Youth and Adolescence, 17, 493-513.Kimlicka, T. M., Cross, T. H., & Tarhai, J. (1983). Acomparison of androgynous, feminine, masculine, andundifferentiated women on self-esteem, bodysatisfaction, and sexual satisfaction. Psychology of Women Quarterly, 7, 291-294.Kimlicka, T. M., Wakefield, J. A., & Goad, N. A. (1982). Sexroles of ideal and opposite sexed persons for college140males and females. Journal of Personality Assesment,46, 519-521.Koff, E., Rierdan, J., & Silverstone, E. (1978). Changes ofrepresentation of body image as a function ofmenarcheal status. Developmental Psychology, 14, 635-632.Korabick, K. (1982). Sex role orientation and impressions: Acomparison of differing genders and sex roles.Personality and Social Psychology Bulletin, 8, 25-30.Lamke, L. K. (1982). The impact of sex-role orientation onself-esteem in early adolescence. Child Development,53, 1530-1535.La Sorsa, V. & Fodor, I.G. (1990). Adolescentdaughter/midlife mother dyad: A new look at separationand self-definition. Psychology of Women Quarterly, 14,593-606.Leaper, C., Hauser, S. T., Kremen, A., Powers, S. I.,Jacobson, A. M., Noam, G. G., Weis-Perry, B., &Follansbee, D.(1989). Adolescent-parent interactions inrelation to adolescents' gender and ego-developmentpathway: A longitudinal study. Journal of Early Adolescence, 9, 335-361.LeCroy, C. W. (1988). Parent-adolescent intimacy: Impact onadolescent functioning. Adolescence, 23, 137-147.141Lerner, R. M., Karabenick, S.A., & Stuart, J. R. (1973).Relations among physical attractiveness, bodyattitudes, & self-concept in male and female collegestudents. Journal of Psychology, 85, 119-129.Lerner, R. M., Orlos, J. B., & Knapp, J. R. (1976). Physicalattractiveness, physical effectiveness, & self conceptin late adolescents. Adolescence, 11, 313-326.Levinson, D. (1978). The seasons of a man's life. New York:Alfred A. Knopf.Lippa, R. (1985). Review of the Bem Sex-Role Inventory. InJ.V. Mitchell,(Ed.), The ninth mental measurements yearbook. Nebraska: University of Nebraska Press.Loevinger, J. (1976). Ego development: Conceptions andtheories. San Fransisco, CA: Josey-Bass.Maccoby, E. E. & Jacklin, C. N. (1974). The psychology ofsex differences. Stanford:Stanford University Press.Macoby, E. E., & Martin, J. A. (1983). Socialization in thecontext of the family: Parent-child interaction. In P.H. Mussen (Ed.). Handbook of Child Psychology, Vol. 4.N.Y.: Wiley & SonsMcCaulay, M., Mintz, L., & Glenn, A. A. (1988). Body image,self esteem, and depression-proneness: Closing thegender gap. Sex Roles, 19, 381-391.142Mendelsen, B. K., & White, D. R. (1985). Development ofself- body esteem in overweight youngsters.Developmental Psychology, 21 ,90-96.Messer, B., & Harter, S.(1986). Manual for the Adult SelfPerception Profile. Colorado: University of Denver.Miller, J. B. (1983). The development of women's sense ofself. Wellsley: Stone Center Working Papers Series.Mintz, L., & Betz, N. E. (1986). Sex differences in thenature, realism, and correlates of body image. Sex Roles, 15, 185-195.Moore, S. M. & Rosenthal, D. A. (1980). Sex roles: Gender,generation, and self-esteem. Australian Psychologist,15, 467-477.Noles, S.W., Cash, T.F., & Winstead, B.A. (1985). Bodyimage, physical attractiveness, and depression. Journal of Consulting and Clinical Psychology, 53, 88-94.Offer, D., Ostrov, E., & Howard, K. I. (1982). Familyperceptions of adolescent self-image. Journal of Youthand Adolescence, 11, 281-291.Payne, F. D. (1985). Review of the Bem Sex-Role Inventory.In J.V.Mitchell, (Ed.), The ninth mental measurements yearbook. Nebraska: University of Nebraska Press.Pedhazur, E. J., & Tetenbaum, T. J. (1979). The Bem Sex-Roleinventory: A theoretical and methodological critique.143Journal of Personality and Social Psychology, 37, 996-1016.Petersen, A. (1988). Adolescent development. Annual Reviewof Psychology, 39, 583-607.Petersen, A. C., Schulenberg, J. E., Abramowitz, R. H.,Offer, D., & Jarcho, H. D. (1984). A self imagequestionnaire for young adolescents: Reliability andvalidity studies. Journal of Youth and Adolescence, 13,93-111.Piers, E. V., & Harris, D. B. (1964). Age and othercorrelates of self-concept in children. Journal of Educational Psychology, 55, 91-95.Rierdan, J., & Koff, E. (1980). Representation of the femalebody by early and late adolescent girls. Journal of Youth and Adolescence, 9, 339-346.Rierdan, J., & Koff., E. (1985). Timimg of menarche andinitial menstrual experience. Journal of Youth andAdolescence, 14, 237-244.Robinson, J.B., & Shaver, P.R. (1973). Measures of Social Psychological Attitudes. Ann Arbor: University ofMichigan press.Rosen, G. M., & Ross, A.D. (1968). Relationship of bodyimage to self-concept. Journal of Consulting and Clinical Psychology, 32, 100.144Rosenbaum, M. (1979). The changing body of the adolescentgirl. In M.Sugar (Ed.), Female adolescent development (pp.234-252). New York: Brunner-Mazel.Rosenberg, M. (1965). Society and the adolescent self-image.Princeton, N.J.: Princeton University Press.Rozin, P., & Fallon, A. (1988). Body image, attitudes toweight, and misperceptions of figure preferences of theopposite sex: A compariison of men and women in twogenerations. Journal of Abnormal Psychology, 97, 342-353.Rubin, L B. (1983). Intimate Strangers: Men and Women Together. N.Y.: Harper & Row.Secord, P., & Jourard, S. (1953). The appraisal of body-cathexis: Body-cathexis and the self. Journal of Consulting and Clinical Psychology, 32, 100.Sethi, A.S., & Bala, N. (1983). Relationship between sex-role orientation and self-esteem in collge women.Psychologia, 26, 124-127.Schilder, P. (1935). The image and appearance of the human body. London: Kegan, Paul, Trench Trubner, & Co.Schontz, F. C. (1969). Perceptual and cognitive aspects of body experience. New York; Academic Press.Sholomskas, D., & Axelrod, R (1986). The influence ofmother-daughter relationships on women's sense of self145and current role choices. Psychology of Women Quarterly, 10, 171-182.Siber, E., & Tippett, J. (1965). Self-esteem: Clinicalassessment and measurement validation. Psychological Reports, 16, 1017-1071.Silverberg, S B., & Steinberg, L. (1987). Adolescentautonomy, parent-adolescent conflict, and parentalwell-being. Journal of Youth and Adolescence, 16, 293-312.Simmons, R. G., Rosenberg, I., & Rosenberg, M. (1973).Disturbance of self-image at adolescence. American Social Review, 38, 553-568.Simmons, R. G., & Rosenberg, I. (1975). Sex, sex-roles, andself-image. Journal of Youth and Adolescence, 4, 229-258.Steiner-Adair, C. (1986), The body politic: Normal femaleadolescent development and the development of eatingdisorders. Journal of the American Academy of Psychoanalysis, 14, 95-114.Strober, M., & Yager, J. (1985). A developmental perspectiveon the treatment of anorexia nervosa in adolescents. InD.M.Garner & P.E.Garfinkel (Eds.), Handbook of psychotherapy for anorexia nervosa and bulimia (pp.363-390). New York: Guilford Press.146Surrey, J. L. (1985). A theory of women's development.Wellesley: Stone Center Working Papers Series.Tapley, D.F., Morris, T. Q., Rowland, L. P., Weiss, R. J.,Subak-Sharpe, G. J., & Goetz, D. M. ( Eds.) (1989). TheColumbia University College of Physicians and Surgeons Complete Home Medical Guide, Revised Edition. N.Y.:Crown Publishing, Inc.Thompson, J. K., & Dolce, J. J. (1989). The discrepancybetween emotional versus rational estimates of bodysize, actual size, and ideal body ratings: Theoreticaland clinical applications. Journal of Clinical Psychology, 45, 473-478.Thorbeke, W., & Grotevant, H. D. (1982). Gender differencesin adolescent interpersonal identity development.Journal of Youth and Adolescence, 11, 479-192.Tobin-Richards, M. H., Boner, A. M., & Petersen, A.0 .(1983). The psychological significance of pubertalchange: Sex differences in perceptions of self duringearly adolescence. In J.Brooks-Gunn & A.C.Petersen(Eds.), Girls at puberty, (pp.127-154). New York:Plenum Press.Van Rackley, J., Warren, S A., & Bird, G.W. (1988).Determinants of body image in women at mid-life.Psychological Reports, 62, 9-10.147Whitely, B. E., Jr. (1983). Sex-role orientation and self-esteem: A critical meta-analytic review. Journal ofPersonality and Social Psychology, 44, 765-778.Whitely, B. E., Jr. (1988). Masculinity, femininity, andself-esteem: A multitrait, multi-method analysis. SexRoles, 18, 419-431.Winstead, B. A., & Cash, T. F. (1984). Reliability and validity of the Body-Self Relations Questionnaire: A new measure of body image. Paper presented at themeeting of the Southeastern Psychological Association,New Orleans, LA.Wooley, S. C., & Wooley, 0. W. (1985). Intensive outpatientand residential treatment for bulimia. In D.M.Garner &P.E.Garfinkel (Eds.), Handbook of psychotherapy for anorexia nervosa and bulimia (pp.391-430). New York:Guilford Press.148APPENDIX ADemographic information: Please answer the followingquestions. Your answers will be held in the strictestconfidence.Mothers 1) Date of birth:2) Height (feet/inches)Weight (pounds)3) Highest level of education completed:elementary schoolhigh schoolcollegeuniversitypost-graduate4) Present employment status:paid full timepaid part timeunemployedhomemakerOccupation:5) Were you employed during the first 5 years of yourdaughter's life?yesno6) If so, were you employed:paid full timepaid part timeOccupation1497) Combined family income:less than $25,000 per year$26,000 - $35,000 per year$36,000 - $45,000 per year$46,000-^$50,000 per yearmore than $50,000 per year8) Who lives in your household? Please indicate numberswhere applicable.motherfatherdaughtersonstepdaughterstepsongrandmothergrandfatherfoster childstepmotherstepfatherother (specify)9) Has your family situation changed significantly in thepast 2-3 years? (eg., divorce, additoin of a family member,lossof a family member)yesnoIf so,describe10) Do you regularly participate in a religiousorganization?yesnoFaith ordenomination:Frequency ofparticipation:11) Do your children regularly participate in a religiousorganization?150yesnoFaith ordenomination:Frequency ofparticipation:12) What is your ethnicheritage?13) What is your daughter's ethnicheritage?14) Are you or is your daughter currently being treated fora mental health concern?ExplainThank you for your participation in the study. If you wouldlike information on the results of the study, please send anote with your name and address to:Sonia Usmiani308-2211 West 2nd Ave.Vancouver, B.C.V6K 1H8I would be happy to send you the results as soon as theybecome available.Yours truly,Sonia Usmiani151APPENDIX BDemographic information: Please answer all questions. Youranswers will be held in the strictest confidence.Daughters 1) Age2) Grade ^3) Height (feet/inches) ^Weight (pounds)4) Do you have any brothers and sisters?yesnoIf so, please indicatenumber of sisters^ ages of sisters^number of brothers ages of brothers5) How old were you when you had your first menstrualperiod?Comments:Thank you for participating in the study. If you would likeinformation on the results of the study, please send a notewith your name and address to:Sonia Usmiani]308-2211 West 2nd. Ave.Vancouver, B.C.V6K 1H8I will be happy to send you the results of the study as soonas they become available.Yours truly,152Sonia UsmianiAPPENDIX CBelow is a list of a number of personality characteristics.We would like you to use those characteristics to describeyourself, that is, we would like you to indicate, on a scaleof 1 to 7, how true of you each of these characteristics is.Please do not leave any of characteristics unmarked.1^2^3^4^5^6^7Never or Usually Sometimes Occasion- Often Usually Alwaysalmost^not^but^ally true true true^ornever true infrequently^ Almosttrue true Alwaystrue153Defend my own beliefsAffectionateConscientiousIndependentSympatheticMoodyAssertiveSensitive to the needsof othersReliableStrong personalityUnderstandingJealousHave leadershipabilitiesEager to soothe hurtfeelingsSecretiveWilling to takerisksWarmAdaptibleDominantTenderConceitedWilling to take astandLove childrenTactful ForcefulCompassionateTruthfulAggressiveGentleConventionalAPPENDIX DAfter carefully reading each of the statements on thefollowing pages, please circle the number which indicateshow well the statement describes you. Each number goes witha category telling whether it describes you VERY WELL (1),WELL (2), FAIRLY WELL (3), NOT QUITE (4), HARDLY (5), NOTAT ALL (6). There are 11 statements. Please respond to allof them. Remember, there are no right or wrong answers.1-Describes me very well^4-Does not quite describe me2-Describes me well^5-Hardly describes me3-Describes me fairly well^6-Does not describe me at all1.I am not satisfied with my weight 1 2 3 4 52.Most of the time I am happy withthe way I look.1 2 3 4 5 63.1n the past year I have been veryworried about my health1 2 3 4 5 64.1 wish that I were in better physicalcondition1 2 3 4 5 65.1 am uncomfortable with the waymy body has developed1 2 3 4 5 66.1 am proud of my body 1 2 3 4 5 67.1 am satisfied with my height 1 2 3 4 5 68.1 frequently feel ugly andunattractive1 2 3 4 5 69.When others look at me, they mustthink that my body did notdevelop well1 2 3 4 5 610.My body developed the way I wouldhave liked it to1 2 3 4 5 611.1 feel strong and healthy 1 2 3 4 5 6154APPENDIX EAfter carefully reading the statements on the followingpage, please circle the number which indicates how well thestatement describes you. Each number has a category tellingwhether it describes you VERY WELL (1), WELL (2), FAIRLYWELL (3), NOT QUITE (4), HARDLY (5), NOT AT ALL (6). Thereare 11 statements. Please respond to all of them. Remember,there are no right or wrong answers.1-Describes me very well^4-Does not quite describe me2-Describes me well^5-Hardly describes me3-Describes me fairly well^6-Does not describe me at all1.1 am not satisfied with my weight 1 2 3 4 5 62.Most of the time I am happy with theway I look1 2 3 4 5 63.1n the past year I have been veryworried about my health1 2 3 4 5 64.1 wish that I were in betterphysical condition1 2 3 4 5 65.1 am uncomfortable with the waymy body is developing1 2 3 4 5 66.1 am proud of my body 1 2 3 4 5 67.1 am satisfied with my height 1 2 3 4 5 68.1 frequently feel ugly andunattractive1 2 3 4 5 69.When others look at me they mustthink that I am poorly developed1 2 3 4 5 610.My body is growing as quickly as 1 2 3 4 5 6I would have liked it to11.1 feel strong and healthy 1 2 3 4 5 6155APPENDIX FAnswer the following questions by circling the number whichbest represents how you feel.Strongly disagree^ 1Disagree^ 2Agree 3Strongly Agree^ 41. On the whole,^I am satisfiedwith myself2. At times I think I am no goodat all3. I feel that I have a numberof good qualities1112223334444. I am able to do things as wellas most people 1 2 3 45. I feel that I do not have muchto be proud of 1 2 3 46. I certainly feel useless at times 1 2 3 47. I feel that I am a person of worth,^1 2 3 4at least on an equal plane with others8. I wish that I could have morerespect for myself 1 2 3 49. All in all,^I am inclined to feelthat I am a failure1 2 3 410.I take a positive attitude towardmyself 1 2 3 4156APPENDIX GSonia Usmiani308-2211 West 2nd. AveVancouver, B.C.V6K 1H8Telephone: 733-6842Dear Volunteer,As a graduate student of counselling psychology at theUniversity of British Columbia, I am interested ininvestigating personality variables and body image in womenand girls. I am including adolescent girls and their mothersin my study. The study is being conducted under thesupervision of Dr. J. Daniluk, who can be reached at 228-5768 to answer any questions you might have.If you participate in the study, you will be asked to fillout four questionnaires on your own time and mail them tome. It will take approximately one hour to answer thequestions. All information on the questionnaires will beheld in the strictest confidence. Through the use of codenumbers, all information will remain anonymous. Your consentto participate in the study will be assumed if you returnthe completed questionnaires to me.If you would like a copyof the results, send a note to me at the address indicatedabove, and I will mail them to you as soon as they becomeavailable.Your participation in the study is entirely voluntary. Youare free to terminate your consent and withdraw yourparticipation at any time. As well, if you have questionsabout your participation in the study, you are free tocontact me or Dr. J. Daniluk at any time.I hope that you will find your participation in the studyinteresting and informative, and that the results of thisstudy will be beneficial both to you and and to other women.Should you have any questions about the study, or want tosign up for participation, please contact me at 733-6842.Thank you for your help.Yours truly,157Sonia UsmianiAPPENDIX HStandardized Scatterplot Diagrams forSIQYA scoresSample I/ Daughters^ Sample 1/ Mothers158'RESID+^ + +Standardized ScatterplotAcross - •PRED^Down^'RESIDOut ++  ^++3 +I^ I2 +I I1 +I^. .^: •II^. i-1 +I II3 OutStandardized ScatterplotAcross^•PRED^DownOut ++^ +^ +^ +3210-2-3Out ^++-3^-2^-1^0^1^2^3 Out-2 +..3^4.Out ++^ + +^ +^ +^ +-3^-2 -1 0 1 2Sample II/ Daughters Sample II/ MothersStandardized ScatterplotAcross^•PRED^Down^Out ++       'RESIDStandardized ScatterplotAcross - 'PRED^Down^ + ^Out ++^ +^ +^ + 'RESID3 + 3 +2 + I^I21 + 10 + I^0 II-1^+ -1I-2^+ -2 I-3^+ -3^Out ++^-3^-2^-1^0 1 2O ut^+4.^-3^-2^-1^03 Out 1 2++3 Out


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items