- Published: January 12, 2022
- Updated: January 12, 2022
- Level: College Admission
- Language: English
- Downloads: 9
Gender Differences in Adolescent Self-Esteem: An Exploration of Domains By: Quatman, Teri, Watson, Cary, The Journal of Genetic Psychology, 00221325, , Vol. 162, Issue 1 Database: E-Journals ABSTRACT. The relationship between gender and global self-esteem in adolescence, while modest, has been well established, with boys consistently scoring higher than girls. In the present study, we sought to understand gender differences in adolescent self-esteem in terms of its component parts. With a relatively large (n = 545) sample of adolescents, drawn from Grades 8, 10, and 12, we specified 8 domains of adolescent self-esteem (personal security, home/parents, peer popularity, academic competence, attractiveness, personal mastery, psychological permeability, and athletic competence) across a number of different instruments and brought them together into a common assessment superstructure. Gender differences as well as the relative contributions of the different domains to overall self-esteem scores were measured. As predicted, boys attained slightly higher global self-esteem scores than girls did, by a difference of . 22 standard deviation units. Contrary to our expectation of more balanced domain effects, boys significantly outperformed girls in 6 of 8 domains, whereas the 2 remaining domains exhibited no significant gender differences. There were no main or interaction effects for grade level. In terms of relative contribution of these domains to global self-esteem for the 2 genders, global self-esteem in boys and girls is predicted in very similar strengths and in the same order of magnitude by identical domains of self-esteem: home/parents, personal security, academic competence, attractiveness, and personal mastery–yielding multiple R2s from . 88 to . 91. Key words: adolescence, gender differences, self-esteem SELF-ESTEEM is associated with a number of important psychological phenomena, both positive and negative. High self-esteem has been associated with productive coping strategies, enhanced motivation, and a positive emotional state (Harter, 1990b). Those with high self-esteem experience an incremental improvement in their quality of life. Low self-esteem, on the other hand, puts an individual more at risk for many emotional and behavioral disorders, such as anxiety, lack of motivation, suicidal behavior, eating disorders, delinquency, conduct disorders, and depression (Harter, 1990b). Low self-esteem has been established as a correlate or antecedent (or both) to depression, and high self-esteem as a protective factor against depression (e. g., Allgood-Merton, Lewinsohn, & Hops, 1990; Battle, 1980; Brage & Meredith, 1994; Harter & Whitesell, 1996; King et al., 1993; Lewinsohn et al., 1994; Lewinsohn, Seeley, & Gotlib, 1997). Self-esteem during the adolescent years appears to undergo a process of metamorphosis. Longitudinal studies of adolescent self-esteem have revealed a decline in self-esteem at age 11, a low between ages 12 and 13 (likely due to shifts in school environment and the onset of puberty), and then gradual, systematic improvements in self-esteem through Grade 12 (Demo & Savin-Williams, 1983; Eccles et al., 1993; McCarthy & Hoge, 1982; Rosenberg, 1986). Although some cross-sectional studies contradict these findings (Brack, Orr, & Ingersoll, 1988; Mullis, Mullis, & Normandin, 1992), adolescence is clearly a pivotal and change-related time in the context of self-esteem. This period of growth and development has a unique transitional nature (Eccles et al., 1993), necessitating adjustments and changes in self-definition that appear to disturb global self-esteem (Wigfield & Eccles, 1994). Adolescence brings with it increased differentiation in self-concept as well as increased cognitive capacity for abstraction and self-reflection (Harter, 1990b). Simmons and Blyth (1987) highlighted an adolescent’s increase in self-determination, peer orientation, self-focus, self-consciousness, concern over opposite-sex relationships, and capacity for abstract cognitive activity. All of these issues–newly salient in adolescence–occur simultaneously with bodily changes at the onset of puberty and ” gender intensification” pressures (Hill & Lynch, 1983), and together they create significant challenges to a young person’s self-esteem. Moreover, the prevalence of symptomatology related to depression increases dramatically between childhood and adulthood (Allgood-Merton et al., 1990). Because self-esteem is thought to be an important contributor to depression, looking at self-esteem in adolescence can illuminate our understanding of depression as well. One consistent piece of the adolescent self-esteem puzzle is that boys outscore girls on global measures of self-esteem. The relationship between gender and self-esteem in adolescence, while modest, has been well established (see Bolognini, Plancherel, Bettschart, & Halfon, 1996; Brage & Meredith, 1994; Chubb, Fertman, & Ross, 1997). O’Brien et al. (1996) reported in a recent meta-analysis of 80 research studies on adolescent self-esteem that boys enjoy slightly higher global self-esteem levels than girls by an average difference of one fifth of a standard deviation (d value = . 20). This small yet statistically significant difference (see Cohen, 1977) becomes more compelling when we consider its power to persist into adulthood, at which time men continue to enjoy the same slightly higher self-esteem than their female peers (Kling, Hyde, Showers, & Buswell, 1999; Major, Barr, Zubek, & Babey, 1999). That these differences occur is clear. Why they occur or in which domains of self-esteem they occur are issues yet to be understood. Although self-esteem in adolescence has typically been measured by a single score of global or overall self-esteem and the validity of global self-esteem as a construct has been well established (Rosenberg, 1979), global self-esteem is in the process of being ” deconstructed” in order to provide answers to such questions. Research on the deconstructed or ” finer-grained” facets of self-esteem is growing, much of it carried out on children of elementary and middle school age. Harter created a measure (the Perceived Competence Scale for Children; 1982) to assess a child’s sense of competence across four different domains: cognitive competence, social competence, physical competence, and an abstract sense of global self-worth, all contributing differentially to a child’s level of self-esteem. Harter found that children do not feel equally competent in every skill domain, and that those areas in which their perceived competence is low wield a negative impact on their overall self-esteem. In a later study of children 8 to 15 years old, Harter (1990a) identified six clearly differentiated self-esteem domains: scholastic competence, athletic competence, peer/social acceptance, behavioral conduct, physical appearance, and global self-worth. She found that the more a child’s importance ratings exceeded his or her competence ratings, the lower the level of overall self-worth (Harter, 1990a). Wigfield and Eccles (1994) also explored the competence beliefs of elementary and middle school students, finding gender-stereotypical results: Boys exceeded girls in their competence ratings for math and sports, and girls exceeded boys in their competence ratings for reading, music, and English. Burnett (1996) replicated those findings in a large sample of 957 students in Grades 3 through 7. In an investigation of academic and physical domains, Burnett found that boys had higher self-esteem than girls in physical abilities and math, whereas girls had higher self-esteem than boys in reading. Studies of the component parts of self-esteem in middle and older adolescents and the influence of the adolescent’s gender are perhaps more piecemeal. In one study extending Harter’s Perceived Competence Scale for Children to work with 17-year-olds, Cairns, McWhirter, Duffy, and Barry (1990) found that boys exceeded girls in the athletic/physical competence domain of self-esteem. They further discovered that cognitive competence was more closely tied to self-esteem in girls than in boys. Allgood-Merton et al. (1990) found that high school girls’ body image affected self-esteem (which, in turn, affected depression scores). Stein, Newcomb, and Bentler (1992), Hollinger (1985), and Mullis and McKinley (1989) showed that self-esteem in both males and females was positively related to measures of masculinity/instrumentality. Moran and Eckenrode (1991) extended the work of Ouellet and Joshi (1986) by demonstrating the link between social support, social stress, and self-esteem among adolescent boys and girls. Knox, Funk, Elliott, and Bush (1996) investigated the relationship of ” possible selves” (categories of self-concept) to global self-esteem and found that females’ self-esteem is related to multiple domains of possible selves, whereas males’ self-esteem is related to only one domain. Researchers have also begun to explore the differential influence of family relationships and support on adolescent girls’ and boys’ self-esteem. Brage and Meredith (1994) found that, relative to the other factors in their study, family strengths (the extent to which families can cope with problems and conflicts) had the strongest total effect on self-esteem. Demo, Small, and Savin-Williams (1987) demonstrated that an adolescent’s self-concept was positively influenced by perceptions of parental support, participation, and communication. Avison and McAlpine (1992) found that an adolescent’s perceptions of both parents as ” caring” were positively associated with feelings of mastery and high self-esteem, for girls more so than for boys, and that perceptions of parental overprotectiveness were associated with depression in girls but not in boys. Other studies (Baumrind, 1991; Hoffman, Ushpiz, & Levy-Schiff, 1988) have demonstrated the positive association between maternal responsiveness/support and adolescents’ feelings of competence and self-esteem. Both ongoing domain specification and further domain integration are needed to understand adolescent self-esteem in general and gender differences in adolescent self-esteem in particular. Whether boys and girls experience the subdomains of self-esteem differently and to what extent those areas affect global self-esteem differentially for boys versus girls are important questions in terms of how psychologists and educators assess, and ultimately intervene in, the self-esteem deficits of adolescents. Furthermore, such knowledge may help us understand the gender differences found in self-evaluations of competence in adolescents, as well as those found in disorders such as anxiety, lack of motivation, suicidal behavior, eating disorders, delinquency, and conduct disorders (e. g., Allgood-Merton et al., 1990; Gjerde, Block, & Block, 1988; Nolen-Hoeksema & Girgus, 1994; Petersen et al., 1993). In addition, gender differences in adolescent self-esteem may shed light on the gender variance in adolescent depression, as a number of studies have found a stronger relationship between self-esteem and depression in girls than in boys (e. g., Allgood-Merton et al., 1990; Bolognini et al., 1996; Gjerde et al., 1988). In one study (Avison & McAlpine, 1992), in fact, self-esteem was the only psychosocial construct that explained the observed gender differences in depressive symptoms. O’Brien et al. (1996) pointed out that the necessary next step in understanding gender differences in self-esteem would be studies of specific facets or domains of self-esteem, as larger gender effects may well exist but may be obscured within global inventories. Harter (1982, 1990b) confirmed a growing consensus that self-esteem is poorly captured by a single score measure that combines evaluations across multiple domains, masking important distinctions that individuals make about their competencies in the different domains of their lives. Harter (1990b) suggested that a broad array of areas should be included in the assessment of adolescent self-esteem–namely, scholastic/academic competence, athletic competence, peer social acceptance, behavioral conduct, physical appearance, close friendships, romantic appeal, job competence, and global self-worth. In a longitudinal study on Swiss young adolescents, Bolognini et al. (1996) used Harter’s questionnaire (Perceived Competence Scale) to investigate age and gender effects in six subscales: scholastic competence, athletic competence, social acceptance, behavioral conduct, appearance, and global self-worth. They found no change over time (age 12 years to 14 years) for global self-worth, athletic competence, and behavioral conduct. Ratings for appearance and social acceptance went down over time, whereas ratings for scholastic competence went up. Gender effects were found for global self-worth, appearance, and athletic competence, with girls rating themselves lower than boys did. In addition, there was an interaction effect for athletic competence: girls’ athletic competence went up slightly, whereas that of boys went down. Following the work of Harter, O’Brien et al., Bolognini et al., and others, we attempted in the present study to identify gender-differentiating subdomains of self-esteem that are embedded in the scales of commonly used global self-esteem measures for adolescents. In this pursuit, we used the Coopersmith Self-Esteem Inventories, School, Adult, and Short forms (Consulting Psychologists Press, 1981) and the Self-Esteem Index (Pro-Ed, Inc., 1990). We were able to locate within these multifaceted scales four of the domains Harter described as important to adolescent self-esteem (global self-esteem, physical appearance, academic competence, and peer social acceptance; Harter, 1990b) and to construct from items within these inventories four other domains that, based on the literature, we predicted would contribute to the gender variance in self-esteem: (a) personal security: the sense of physical and psychological safety experienced by the adolescent; (b) personal mastery: the individual’s sense of instrumentality or self-efficacy; (c) psychological permeability: how thick- or thin-skinned the individual is; how vulnerable to and affected by sources of stress the individual is; and (d) family/home life: how supported the adolescent feels in the context of parents and home. The remaining domain, athletic competence, was not examined in the source inventories we used in this study, despite its strong conceptual link to gender differences in self-esteem (Cairns et al., 1990), so we constructed a brief scale from other sources. One domain suggested by Harter, behavioral conduct, was used by Bolognini et al. (1996) and was the only domain in that study to show neither a gender effect nor an age effect. We thus chose to exclude behavioral conduct from the domains we investigated. In the present study, we sought to understand global self-esteem gender differences in adolescence in terms of their component parts. (We use the term gender in this work, as opposed to sex, to emphasize the social and psychological constructs of male and female identity and the differences found therein, rather than the more narrowly defined construct of sex and biological differences.) We were able to access a relatively large sample of adolescents, drawn from Grades 8-12, and to assemble into one composite measure a number of those domains most closely associated in other studies with adolescent self-esteem. This in turn allowed us to estimate the relative contributions of these domains to global self-esteem in both boys and girls. Moreover, the use of existing global self-esteem instruments allowed us to glean information on more discrete areas of self-esteem from sources that already exist and that can be easily accessed by others. The ability to understand in finer grain the gender differences in self-esteem that emerge during adolescence allows us to address more specifically the areas of low self-esteem in both adolescent boys and girls. In this context, then, our goals for this study can be summarized as follows: 1. To specify important gender-differentiating domains of adolescent self-esteem, 2. To bring those domains together into a common assessment superstructure, 3. To measure the differential performance of the two genders on the domains of self-esteem so assembled, 4. To estimate the relative contribution of the different domains to overall self-esteem scores across the two genders. Method Participants The participants were 545 adolescents in 8th (n = 274), 10th (n = 157), and 12th grades (n = 114). They were drawn from one elementary school (1st-8th), two middle schools (6th-8th), and two high schools (9th-12th) in the San Francisco Bay area. Each participant completed a 90-min battery of measures selected to elicit a wide range of information about psychological, social, and demographic factors. The data for this study were collected as part of a larger program of research on factors influencing adolescent achievement. Students were administered questionnaires in their classrooms (25-35students) by two research assistants per classroom, typically during two classroom sessions, separated by 1 to 5 school days. Parent and student permission for student participation was solicited through a direct mailing and classroom-based follow-up. Parents and students both provided signed consent for student participation. The overall participation rate was 88%. The sample was evenly divided by gender (50% male and 50% female). The students ranged in age from 12 to 18 years. The sample was ethnically diverse: 63% Caucasian, 18. 8% Asian or Pacific Islander, 10% Latino, 2. 6% African American, and 5. 4% other. Eighty-three percent reported coming from intact or ” one household” families; 17% from ” more than one household.” Socioeconomic status of the families ranged from lower-middle to upper-middle class. The sample contained a higher than usual number of high-achieving adolescents due to the sampling requirements of the overall study: 12% of the students tested were rated by their teachers as superior (98th percentile); 25% of the students were rated by their teachers as high achieving (91st-98th percentile); 27% were rated by their teachers as moderately high achieving (76th-90th percentile); 25% were rated by their teachers as average achieving (26th-75th percentile); and 11% were rated by their teachers as below average achieving (0-25th percentile). Description of Measures The questionnaires included a brief demographic profile and 14 self-report, true-false, and multiple-choice instruments. Instruments were varied in order of presentation within each student packet. Packets were identified by research code number only; students were assured of their anonymity to both researchers and school personnel. Measures included were the Coopersmith Self-Esteem Inventories, School and Adult forms (Consulting Psychologists Press, 1981); the Self-Esteem Index (SEI; Pro-Ed, Inc., 1990); and the Beck Depression Inventory (BDI; Beck & Steer, 1987). Table 1 presents a listing of each domain, the associated measure, and the sources from which they were drawn. Table 2 presents scale and subscale correlations. For a global measure of self-esteem, no single inventory provided a unified instrument (a) appropriate for youngsters of middle school age through the end of high school and (b) sufficiently broad in content to address the domains of interest. We therefore made a composite of two inventories for our analyses: the Coopersmith Self-Esteem Inventories (Consulting Psychologists Press, 1981) and the Self-Esteem Index (Pro-Ed, Inc., 1990). Global Self-Esteem Composite. This 105-item composite scale represented the marriage of the study’s two main self-esteem instruments: 25 items were from the Coopersmith Self-Esteem Inventories, School, Adult, and Short forms (Consulting Psychologists Press, 1981), and 80 items were from the Self-Esteem Index (Pro-Ed, Inc., 1990). The scores were standardized, weighted equally, and averaged together to yield a global self-esteem composite score. Our analyses yielded a Cronbach alpha coefficient of . 94 for the 105-item scale. The scales that make up this composite score are described as follows. 1. The Coopersmith Self-Esteem Inventories, School, Adult, and Short forms (Consulting Psychologists Press, 1981): Two forms of the Coopersmith Self-Esteem Inventories were used in this work, the form dependent on the age of the respondent. The School Short form was administered to adolescents younger than 16 years of age. It consisted of 25 items (e. g., ” Things usually don’t bother me;” ” I often wish I were someone else”) to which students responded like me or unlike me. This scale generated no subscales. This form was correlated with the full length (58-item) School form at . 86 (Coopersmith, 1967). Alpha coefficients for the School Short form ranged from . 78 to . 85. The Adult form was administered to adolescents 16 years or older. It consisted of the same 25 items that appeared in the School Short form, with several quite minor wording changes to adapt the inventory for older respondents (e. g., ” family” is substituted for ” parents” in 4 items; ” group” and ” work” are substituted for ” class” and ” school”). The correlations of total scores on the School Short form and the Adult form exceeded . 80 for three samples of high school and college students. Internal consistency scores for the Adult form ranged from . 78 to . 85 (Consulting Psychologists Press, 1981). The alpha coefficient for the combined School Short and Adult forms was . 81. 2. The Self-Esteem Index (Pro-Ed, Inc., 1990): This index is used with students 8 through 19 years old and consists of 80 items scored on a scale ranging from always true (1) to always false (4). The Self-Esteem Index generates four 20-item subscales, three of which were used in this study: Personal Security (a = . 81), Peer Popularity (a = . 79), and Academic Competence (a = . 87). In this study, the SEI was correlated at . 66 with the 58-item Coopersmith Total Self Score. Subscale intercorrelations ranged from . 42 to . 56 (see Table 1). Domains of self-esteem. Eight domains of self-esteem were identified for this study. For all but one domain, we used one measure per domain. Because a number of domains were measured in parallel by both self-esteem inventories, and given the focus of the present study, we chose in each case the subscale that included the broadest age spectrum available. (See Table 1 for a listing of each domain and the associated measure.) The first four of the eight measures that follow are existing subscales contained within the global self-esteem inventories used for this study; the remaining four were constructed for this study. 1. Personal Security Scale: The Self-Esteem Index (Pro-Ed, Inc., 1990) contains 20 items that pertain to the adolescent’s overall feelings of anxiety, guilt, and shame concerning real or imagined transgressions. It asks about fears and phobias; physiological or emotional indicators of stress; perceptions of personal safety; as well as general health, maturity, comfort, and acceptance within the peer group. Examples are ” I get a lot of headaches and stomachaches,” ” I am often afraid,” ” I often feel ashamed of myself,” and ” I never feel like I’m part of the group.” Our results yielded a Cronbach alpha coefficient of . 81 for the 20-item Personal Security Scale. 2. Home/Parents Scale: Because the Coopersmith Self-Esteem Inventory School form and its parallel Adult form (Consulting Psychologists Press, 1981) overlap in item content on 6 of 8 Home Scale items, we combined them to create a Home/Parents composite that included the responses of the full range of students tested (Grades 8 through 12). This 6-item scale pertained to the adolescent’s experience in his or her home. Items tapped the adolescent’s perceptions of being understood and paid attention to; having his or her feelings considered by parents; whether he or she felt pushed or burdened by parental expectations; and whether home was a place of fun or upset, or a place the adolescent wished to leave. Cronbach alpha coefficients for the resulting composite were . 75 for the 6-item composite. 3. Peer Popularity Scale: The Self-Esteem Index Peer Popularity Subscale (Pro-Ed, Inc., 1990) contains 20 items. Examples are ” I include other people in my plans,” ” My friends don’t have much confidence in me,” and ” I’m usually the last one to be chosen for a game.” Our results yielded a Cronbach alpha coefficient of . 81 for the 20-item Peer Popularity Scale. 4. a. Academic Competence Scale: The Self-Esteem Index Academic Competence Subscale (Pro-Ed, Inc., 1990) contains 20 items. It is concerned with individuals’ perceptions of their school performance, their interest in and desire to excel at academic activities, the value they attach to intellectual achievement, and the support they feel from their teachers. Examples include ” I like going to school,” ” Most of my teachers are pretty fair,” and ” It’s fun to learn new things.” Our results yielded a Cronbach alpha coefficient of . 87 for the 20-item Academic Competence Scale. 5. b. Academic Conscientiousness Scale: This composite was excised from the Academic Competence Scale as an experimental attempt to capture the concepts of conscientiousness and compliance as they relate to the school arena. Examples of the 6 items that tapped this subdomain include ” I am a hard and steady worker at school,” and ” I’m pretty good about doing my homework on time.” This 6-item subscale yielded a Cronbach alpha coefficient of . 70. This scale was not meant to be represented as either conceptually or empirically distinct from the Academic Competence Scale, but was meant simply to focus attention on a part of that scale that may differentiate girls from boys. The correlation between this scale and the Academic Competence Scale was . 87. 6. Attractiveness Scale: The attractiveness/physical appearance composite was an author-generated scale drawn from the sources listed in Table 1. It consisted of 4 items. Examples include ” I’m as nice looking as most other kids,” and ” I believe I look ugly.” Our results yielded a Cronbach alpha coefficient of . 80 for the 4-item subscale. 7. Personal Mastery Composite: This composite was a 9-item author-generated scale with items that focus on the theme of instrumentality and being sure of oneself–for example, ” I’m pretty sure of myself,” ” I’m a leader in most of the games that my friends play,” ” I don’t have trouble making up my mind,” and ” I’m afraid of making mistakes.” Scale sources are listed in Table 1. Items were standardized and averaged. Our results yielded a Cronbach alpha coefficient of . 74 for the 9-item subscale. In other studies, masculinity has been used as an oblique measure of personal mastery or instrumentality (Hollinger, 1985; Stein et al., 1992); our subscale was correlated with the Bem Sex Role Masculinity Inventory at . 43. 8. Psychological Permeability Composite: The Psychological Permeability Composite is a 5-item author-generated scale (sources are listed in Table 1). Items focus on the question of to what extent one is vulnerable to circumstances and relationships that threaten the self–for example, ” I get upset easily when I am scolded,” ” I have nightmares almost every night,” and ” Things usually don’t bother me.” Higher scores indicate poor adjustment. Our results yielded a Cronbach alpha coefficient of . 78 for the 5-item subscale. 9. Athletic Competence Subscale: Despite its conceptual importance, this domain was not represented at all in either of our source self-esteem inventories. The subscale, therefore, is experimental, including only 2 author-generated items: ” I try to achieve in sports” (scored 1-7) and ” How athletic I am” (scored 1-7). Items were standardized and averaged. This scale was retained for its potential value to further research. Our results yielded a Cronbach alpha coefficient of . 80 for the 2-item subscale. Depression. The Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979) is a 21-item instrument designed to assess the severity of depression in adolescents and adults. The symptoms and attitudes associated with depression are rated on a 4-point scale, ranging from 0 (absence of symptom) to 3 (severe presence of symptom). Scores of 10-16 indicate mild depression; 17-29, moderate depression; 30-63, severe depression. Beck, Steer, and Garbin (1988) found a mean correlation of . 72 between clinical ratings of depression and the BDI for psychiatric patients across a variety of studies, and a mean correlation of . 60 between the BDI and clinical ratings for nonpsychiatric subjects. Results On the basis of the moderate gender differences found in global measures of adolescent self-esteem, we had anticipated that boys would exceed girls in our global measures of self-esteem. At the level of self-esteem domains, however, we had anticipated a balancing of effects, with boys maintaining moderately higher self-esteem scores in such domains as personal mastery and athletic competence and lower scores in psychological permeability, and girls achieving higher self-esteem scores in other domains, such as home/parents and peer popularity. To test these hypotheses, we carried out two-way analyses of variance (ANOVAs) with grade (3 levels) and gender as the two factors. The results appear, as standardized z scores for ease of comparison, in Table 3. As predicted, boys achieved higher global self-esteem scores than girls did. In addition, boys scored significantly higher than girls in six of the eight domains, whereas the two remaining domains exhibited no significant differences between the genders (see Table 3), contrary to our hypotheses, in which we expected some balancing of domain effects. (Lower scores in psychological permeability indicate good adjustment. In this domain, boys outperformed girls by achieving lower scores.) Girls did not exceed boys in any major domain of self-esteem studied. They did, however, achieve significantly higher scores in the exploratory sub-domain of academic conscientiousness. This achievement, though, appeared to be of only modest value in terms of contribution to overall feelings about the self (r = . 56, p < . 001). Grade effects were absent in all domains, and there were no gender by grade interactions. Global Self-Esteem Boys exhibited slightly higher global self-esteem than girls did (Global Self-Esteem Composite, p < . 01), in accordance with findings of other researchers. The average magnitude of this difference was roughly one fifth of a standard deviation (z = . 22; see Table 3), replicating the findings of O'Brien et al. (1996). Self-Esteem Domains Personal security. Boys (8th through 12th grades) enjoyed a higher perception of personal security than girls did (p < . 05). They appeared to suffer less from anxiety, guilt, and shame than girls did and reported fewer physiological manifestations of such emotions as well. Girls were more likely than boys to acknowledge physiological symptoms (" I get a lot of headaches and stomachaches"; p < . 0001) and psychological symptoms of emotional disturbance (" I have nightmares most every night"; p < . 01). Home/parents. Perhaps the most striking set of differences was in the domain of family and home life, in which girls reported significantly less satisfaction with their home life and family than boys did (p < . 01). Based on specific items, girls said there were many times when they would have liked to leave home (p < . 005) and that no one paid much attention to them at home (p < . 01). Girls also were less likely than boys to say that their parents understood them (p < . 001). Perception of peer popularity. There were no significant differences between the perceptions of peer popularity of boys and girls. Academics. In overall perception of academic competence, boys and girls were alike. In terms of academic conscientiousness, however, girls saw themselves as more conscientious and more cooperative than boys did in the academic realm (p < . 01). At the item level, girls reported better behavior in school (p < . 0001). Girls were more likely than boys to say that they were hard workers (p < . 005) and that they finished work on time (p < . 005). However, their greater efforts in this realm did not appear to reap greater rewards; in response to an author-generated item not included in the aforementioned subscales, boys in this sample reported being more satisfied than girls with how smart they were (p < . 05). Attractiveness. In general, boys outscored girls on the dimension of attractiveness/physical appearance (p < . 01). Both genders reported (in items not included in any composite) being actively involved in becoming more attractive (girls: " I do what I can to lose weight and be thin," p < . 0001; boys: " I try to lift weights and bulk up," p < . 0001). Based on individual items, boys were more satisfied with how attractive they were (p < . 001), and girls were significantly more likely to say that they " believe they look ugly" (p < . 0001). Personal mastery. According to the Personal Mastery Composite, boys enjoyed a stronger sense of personal mastery than girls did (p < . 0001). It would be expected that boys would score higher than girls on such stereotypical aspects of gendered attitudes and behaviors as leadership and risk-taking. But analyses of individual items from the Personal Mastery Composite indicated, in addition, that boys were less afraid than girls were to make mistakes (p < . 0005) and to fail (p < . 0001). Boys were more sure of themselves (p < . 005) and were more apt to feel like a leader among their peers (p < . 005). Psychological reactivity/permeability. Girls were clearly more likely than boys to report being affected by their emotions (p < . 0001). Although no other scale tapped this aspect of adolescent experience, this composite yielded the highest F value of any measure used in our study, F(1, 536) = 38. 6. Girls reported getting upset in response to being scolded (p < . 0001), and they felt that other kids thought they were crybabies (p < . 005). Girls also reported more psychosomatic symptoms, including nightmares (p < . 005), headaches, and stomachaches (p < . 0001). Athletics. Boys felt more satisfied with their athletic competence (p < . 0001) and were more likely than girls to say they tried to achieve in sports (p < . 0001). In this domain, girls made less effort and were less satisfied than boys were. Interestingly, this was the only domain in which gender by grade interaction effects approached significance. For both genders, but more prominently for girls-interaction effect: F(1, 520) = 2. 6, p = . 08 athletic competence was felt to be lower in 12th grade than in either 8th or 10th grades. In summary, male adolescents scored more positively than female adolescents in global self-esteem and in six of the eight self-esteem domains. Girls outscored boys in academic conscientiousness only, a subdomain of academic competence. The only subscales in which no significant gender differences were found were peer popularity and overall perception of academic competence. Linking Self-Esteem Domains to Overall Self-Esteem: Comparing the Genders To investigate the strength of domain contributions to overall self-esteem in the two genders, we further analyzed the data in a series of stepwise regression analyses in which several outcome variables were considered: (a) the Global Self-Esteem Composite; (b) the Global Self-Esteem Composite, split by the gender of the respondent; and (c) the Beck Depression Inventory. In each of these analyses, eight self-esteem domain variables were entered as potential predictors: personal security, home/parents, peer popularity, attractiveness, academic competence, personal mastery, psychological permeability, and athleticism. In addition, where appropriate in two of the three analyses, gender was also entered as a predictor. The order of entry was allowed to vary according to the relative explanatory power of the variables. The results were highly similar across the three sets of outcomes related to the Global Self-Esteem Composite (see Table 4). Whether the sample was taken as a whole or split by gender, the same five variables entered as significant predictors of global self-esteem-home/parents, personal security, academic competence, attractiveness, and personal mastery-yielded multiple Rs from . 88 to . 91. In the full sample, home/parents was the most potent predictor of global self-esteem (as signified by the beta weights). This was followed by personal security, personal mastery, and then, more weakly, by academic competence and attractiveness. When the genders were considered separately, regression results were remarkably similar to one another. For both genders, home/parents was the strongest predictor of overall self-esteem, followed by personal security. For both genders, these first two variables accounted for more than two thirds of the variance in global self-esteem. The remaining three predictors to be entered in the model were, for both genders, academic competence, attractiveness, and personal mastery. For girls, the order of entry was academics, followed by attractiveness and personal mastery; for boys, it was attractiveness, followed by academics and personal mastery. The prediction equation looked somewhat different when depression was identified as the outcome variable. In that case, feeling unattractive most strongly predicted depression, followed by unhappiness with home/parents, then psychological permeability, lack of personal mastery, being female, peer popularity, and the lack of academic competence. In summary, global self-esteem in both adolescent boys and girls was associated with a positive home and family life, a strong sense of personal security, self-perceived academic competence, attractiveness, and personal mastery. Each of these variables made an independent contribution to the prediction of global self-esteem for adolescents. Discussion Two concerns have been the foci of this research: (a) whether differences in global self-esteem between male and female adolescents can be understood more productively in terms of their constituent domains, and (b) to what extent those domains predict self-esteem differently for the two genders. These questions will be taken up in order. Gender differences in global or general self-esteem have been replicated in this study. Boys exhibited slightly higher global self-esteem than girls did (Global Self-Esteem Composite: p < . 01). The magnitude of the difference (z score = . 22; see Table 3) was similar to that found by O'Brien et al. (1996): an average of one fifth of a standard deviation across 80 studies of gender differences in global self-esteem. Like other researchers, we found that boys enjoy slightly higher self-esteem than girls do. What is a new and disturbing finding in this study is that those differences appear to be ubiquitously distributed across the multiple domains of self-esteem examined in this process. Adolescent girls scored significantly lower than their male counterparts in all but two of the domains of self-esteem identified in this study, and exceeded them in none. Boys strongly exceeded girls in a number of self-esteem domains, notably, general self-worth, athletic competence, personal mastery, and home/parents. They were also markedly less psychologically permeable/reactive than girls were. These domains bespeak a sense of " at-homeness" in the world-a sense in which, starting in their own home environments and moving out from there, boys feel relatively confident and masterful as they encounter life's challenges and demands. They are not undone by adversity, but rather rise to the physical, instrumental, and psychological challenges. In contrast, girls feel significantly less confident and masterful, and more psychologically vulnerable across the full grade range sampled (8th through 12th). The realm of athletic performance has become a focus of intense interest for developing adolescents, as opportunities for girls to cultivate and develop athleticism have burgeoned during the past decade. Attitudes toward athleticism in women have shifted in suit, such that both boys and girls appear to value participation in this formerly male domain (Quatman, Watson, Sampson, & Robinson, 2001). Despite the opening of sports to female participants, girls are behind boys in their satisfaction with how athletic they are and perceive themselves as exerting less effort to achieve in sports than boys do. This result may be linked to girls' comparing themselves with boys in terms of the clear physical increment in power, strength, and large muscle coordination enjoyed by their adolescent male counterparts. But it also may well be linked to girls' unconscious rejection of their own bodies (see discussion that follows). In terms of personal mastery, girls of pre-adolescent ages are given full encouragement to pursue personal mastery of virtually any domain (Pipher, 1994). As those same girls move into adolescence, they must begin to calibrate their sense of personal mastery to their desire to be well received by both boys and other girls. They must at some level make inferences about what the social ethos will bear in terms of avenues and expressions of competence. They must answer internal and perhaps unconscious questions, such as which pursuits are considered " okay" for adolescent girls and which are not. If I succeed at this particular pursuit and appear consciously to enjoy that success, will that confidence threaten boys? Will it put off other girls? Will teachers and other adults find such attitudes " unfeminine"? In the midst of such concerns, it is quite understandable that adolescent girls would develop attributional styles that place the fulcrum of success somewhere outside themselves and their abilities. Not surprisingly, attributions of success to luck or to the " ease" of the task are much more common in adolescent girls than in adolescent boys (e. g., DeMoss, Milich, & DeMers, 1993; Gjerde et al., 1988). The question as to why girls reported significantly less satisfaction with their home life and family than boys did across virtually all items and subscales is an intriguing one. Demo et al. (1987) found that male self-esteem was more strongly related to family relationships than was female self-esteem-a result replicated in this study. What may be at work in this domain are gender differences in a necessary developmental task of adolescence. Despite their stronger identification with their own mothers, which would seem to lead girls to have a less alienating experience of family life than boys have, adolescent girls are faced with a more difficult task in separating from home and mother than boys. An adolescent girl, then, must in essence create enough differentiating momentum and aggression to break free of the formerly comfortable orbit defined by home and parents. To do so, girls often use the vehicle of devaluing their parents, especially their mothers, perceiving themselves as different from and poorly understood by parents. The genders differed strongly on items related to psychological permeability, even though the correlation between this domain and global self-esteem was modest compared with the other domains (see Table 2). The girls appeared to have " thinner skin" than the boys and to be bothered more by physical manifestations of psychological turmoil. This result may reflect what Chodorow (1978) described as a permeability related to the psychic roots of the female experience: because girls can enjoy a merger and identification with their mothers longer than boys can, girls develop more permeable boundaries in relationships and a greater investment in calibrating decisions in terms of relationship-sometimes at the expense of the self. More compelling, however, is the perspective from French psychoanalyst Jacques Lacan (1998), who pointed out that the linguistic/symbolic system awaiting the developing boy/man is vastly more descriptive of the male experience than the one available to the developing girl/woman. As the developing girl/woman searches for vehicles to carry her subjectivity, she faces the poverty of symbolic language to represent her experience. Given this lack, aspects of female experience remain trapped in them as physical symptomatology, a phenomenon that, in its extreme manifestation, Freud referred to as hysteria. In the domains of physical attractiveness and personal security, boys exceeded girls moderately, F(1, 428) = 9. 1, p < . 01, and F(1, 472) = 6. 1, p < . 05. Certainly, there is a cultural imbalance in terms of the pressure felt to be physically attractive, with greater pressure felt by girls and women. The results of this analysis suggest that despite that pressure, or perhaps because of it, adolescent girls are less confident in their physical attractiveness than boys are. Personal security in this study represents a broad psychological construct denoting overall feelings of well-being. Studies preceding this work have found these domains to be linked quite intimately (e. g., Harter, 1990a, 1990b; Lewis & Brooks-Gunn, 1979; Simmons & Blyth, 1987), based presumably on the effects of girls' receiving and internalizing ambiguous signals about their emerging physical development. Common notions of attractiveness in young women center on an anorexic ideal that bears more resemblance to the pre-adolescent than to the post-adolescent female. The biologically impelled acquisition of increased fatty deposits to legs, buttocks, and stomach is unwelcome by most developing female adolescents and, in many, becomes the impetus for eating disorders (Pipher, 1994). Moreover, in terms of sexual development, girls, in contrast to boys, endure a relatively mute and less incontrovertible passage into their enhanced sexual responsiveness, whereas their male counterparts often speak explicitly of topics related to sexual maturation. Girls are left questioning whether the transition to womanhood will bode well for them and are moved along in that transition, whether they welcome it or not. Given the stress on the fabric of self exerted by these factors, it is little wonder that girls report lower feelings of liking of self and attractiveness than do their male counterparts. Two domains that evidenced no significant differences between the genders were peer popularity and academic competence. The male mean for peer popularity exceeded the female mean by roughly . 2 of a standard deviation, although it did not attain conventional levels of significance. However, the trend is certainly in the same direction as the results across other domains-boys exceeding girls. The strongest single-item difference between girls and boys was in response to the item " being popular is not important in the long run," with girls endorsing this item more intensely than boys (p < . 001). This item result is an interesting one, given the collective social orientation of girls versus the more individualistic social orientation of boys (Kimmel & Rudolph, 1998). It may be that boys are socialized into a better understanding than girls of the long-range benefits of networks. Certainly in terms of such ultimate structures as the adult workplace, being popular or at least being liked may mean being chosen over others for jobs, promotions, and leadership roles. Academic competence stands out in this study as one of the few areas in which boys do not feel better about themselves than girls do. As such, this domain invites a closer look. Our 20-item academic competence scale sampled individuals' perceptions of their school performance, their interest in and desire to excel at academic activities, the value they attached to intellectual achievement, and the support they felt from their teachers. It is important to note that in our sample, despite equal-sized gender groups and no systematic gender sampling bias, the girls outperformed the boys academically (mean grade point average = 3. 01 for girls and 2. 77 for boys, p < . 005). All things being equal, the girls in our sample should have enjoyed higher academic self-esteem than the boys did, yet the overall mean for girls exceeded the mean for boys by only . 15 standard deviation units, with only trend-level significance (p < . 13). Academics is one arena in which girls receive frequent, systematic, objective, unambiguous feedback about their performance. It is an arena that does not undergo substantive change during adolescence, only incremental change. In principle, girls should be able to preserve a sense of continuity of academic self-esteem from pre-adolescence into adolescence. But the girls in our sample did not enjoy higher academic self-esteem than the boys did. Interesting, moreover, is the fact that when these same girls were asked to rate on a 1-7 scale their satisfaction with how smart they were, as a group they expressed lower satisfaction ratings than their male counterparts did-mean satisfaction = 5. 33 for boys and 5. 11 for girls; F(1, 494) = 4. 4, p < . 05. Cairns et al. (1990) suggested that the close link between cognitive competence and self-esteem in girls (more so than in boys) may indicate that girls underrate their academic competence whereas boys overrate it, which may explain the results seen for satisfaction with academic performance and boys' and girls' equivalent levels of academic self-esteem. In the service of further understanding this important area of gender-mediated difference within domains, we parsed academic competence into a finer-grained subdomain, one that inquired only about issues of school-related conscientiousness/compliance. In this arena, girls squarely outperformed boys (p < . 01). It appears that girls play by the rules in the context of school assignments and classroom-related behavior. They take their job as students more seriously than do their male counterparts. Our data appear to capture girls working more diligently and more successfully in their academics but, as mentioned earlier, enjoying less sense of satisfaction than might accompany their more arduous academic labors. Overall, the imbalance in the results in favor of boys' self-esteem, with effects of variable magnitude, was both unanticipated and perplexing. Our tacit aim was to discover the differential distribution of self-esteem factors across the two genders. What emerged, instead, was a generally poor report card for girls' versus boys' self-perceptions across six of the eight domains studied. Several theories compete for the explanation of these results. The most prominent come from the research of Nolen-Hoeksema and Girgus (1994), who pointed to a certain ruminative cognitive style in girls. Akin to the style observed in obsessive-compulsive individuals, Nolen-Hoeksema and Girgus argued that girls rehearse cognitive content in their minds over and over, making negative aspects more and more salient. Research on attributional styles and learned helplessness in young men versus young women reveals inconsistent and mixed results (Nolen-Hoeksema & Girgus, 1994). Although some studies support the existence of more pessimistic attributional styles in prepubescent girls (e. g., Dweck & Bush, 1976; Frey & Ruble, 1987; Parsons, Meece, Adler, & Kaczala, 1982), it seems that these traits are not specific to females, as they are correlated with depressive symptoms in both boys and girls (Nolen-Hoeksema, Girgus, & Seligman, 1991). Pipher (1994) suggested that the domains of self-expression open to girls of pre-adolescent ages begin to close down to adolescent girls. Attendant to this closing are both the truncating of options and the depression that accompanies any loss. Compared with boys, adolescent girls must deal with mixed social messages about a number of the domains measured in the present study. We turn finally to our question of to what extent the domains studied predict self-esteem differently for the two genders. Is one domain vastly more important to address, or does one domain affect girls more than boys? The remarkable result in this set of comparisons is the absence of gender-mediated differences in terms of either the weight or the priority of the various self-esteem domains examined. Considered from both regression and correlational analysis perspectives, an array of items strongly influence self-esteem in both genders: parents and home life, personal security, academic competence, and personal mastery. Two rather surprising results were the low priority assigned to athleticism for both boys and girls-clearly the last runner in regression analyses-and the assignment of more self-esteem-related importance to personal appearance by boys than by girls. But clearly the salient message here is the agreement between the genders in terms of what influences self-esteem in adolescence. One methodological concern for any study involving self-report data is the possibility of a social desirability response bias (i. e., the desire to present a positive self-image in response to self-report questions). An immediate reaction to our results is to wonder if the boys in our sample, following internal gender-stereotypical pressures, underreported negative feelings when providing data, thus creating a gender difference in self-esteem domains across the board. Many researchers have considered the possibility of a gender-specific self-presentational bias already and have concluded that it is unlikely to have influenced their results (e. g., Beyer & Bowden, 1997; King & Buchwald, 1982; Koenig, Isaacs, & Schwartz, 1994; Nolen-Hoeksema et al., 1991; Petersen et al., 1993). Given the anonymous nature of our self-report data and the openness allowed by such a collection method, we also conclude that a gender-specific response bias has not influenced our results. The possibility of a gender response bias cannot be completely eliminated, but based on current data, such an artifact seems an unlikely main effect. In summary, then, boys exceeded girls by small but significant amounts in six of the eight self-esteem-related domains identified. Girls exceeded boys in none of the domains. Global self-esteem in males and females is predicted in very similar strengths by identical domains of self-esteem. The results of the present study suggest the utility of such finer-grained and gender-sensitive approaches to the study of adolescent self-esteem. Given the link between self-esteem and such other important aspects of the adolescent experience as quality of life, motivation, eating disorders, depression, and resiliency, gender differences should continue to be studied and ultimately targeted in gender-specific prevention programming. This research was supported by an Arthur Vining Davis Junior Faculty Fellowship at Santa Clara University. The authors gratefully acknowledge Shirley Feldman for critical reading of the manuscript, Diane Wydler for data analytic consultation, and Cindi Robinson for assistance with organization and administration of the project. Address correspondence to Teri Quatman, Department of Counseling Psychology and Education, 226 Bannan Hall, Santa Clara University, Santa Clara, CA 95053; send e-mail to tquatman@scu. edu. TABLE 1 Global Measures and Domain Measures of Self-Esteem and Their Sources Legend for Chart: A - Measure B - Number of items C - alpha D - Source A B C D Global self-esteem Global self-esteem composite 105 . 94 Coopersmith Self-Esteem Inventories(a) (25) Self-Esteem Index (SEI) (80) Self-esteem domains Personal security 20 . 81 Perception of Personal Security Subscale (SEI) Home/Parents 6 . 75 Home/Parents Subscale, School and Adult forms(a) (Coopersmith) Peer popularity 20 . 81 Perception of Peer Popularity Subscale (SEI) Academic competence 20 . 87 Perception of Academic Competence Subscale (SEI) Academic conscientiousness 6 . 70 composite Perception of Academic Competence Subscale items Attractiveness 4 . 80 Coopersmith Self-Esteem Inventories(a) (1) Perception of Peer Popularity Subscale (SEI) (1) Beck Depression Inventory (1) Author-generated item (1) Personal mastery 9 . 74 Coopersmith Self-Esteem Inventories(a) (2) Perception of Peer Popularity Subscale (SEI) (2) Author-generated items (5) Psychological permeability 5 . 78 Perception of Personal Security Subscale (SEI) (3) General Self Subscale, School and Adult forms(a) (Coopersmith) (2) Athletic competence 2 . 80 Author-generated items (2) (a) School and Adult forms of Coopersmith Self-Esteem Inventories contain 25 common items that constitute a " Short" version of the School form. TABLE 2 Correlations Among Measures of Global Self-Esteem (SE), Self-Esteem Domains, and Depression Legend for Chart: A - Measure B - Global SE composite(a) C - Personal security D - Home/Parents E - Peer popularity F - Attractiveness G - Academic competence H - Academic conscientiousness I - Personal mastery J - Psychological permeability K - Athletic competence L - Beck Depression Inventory A B C D E F G H I J K L Global SE composite 1. 00 . 71 . 72 . 66 . 44 . 66 . 56 . 62 -. 55 . 19 -. 65 Personal security 1. 00 . 32 . 59 . 28 . 45 . 33 . 48 -. 60 . 13 -. 39 Home/Parents 1. 00 . 21 . 11 . 37 . 34 . 36 -. 35 . 16 -. 49 Peer popularity 1. 00 . 41 . 45 . 37 . 48 -. 36 . 17 -. 31 Attractiveness 1. 00 . 18 . 19 . 32 -. 18 . 21 -. 47 Academic competence 1. 00 . 90 . 25 -. 34 . 09 -. 36 Academic conscientiousness 1. 00 . 21 -. 21 . 13 -. 29 Personal mastery 1. 00 -. 46 . 15 -. 47 Psychological permeability 1. 00 -. 12 . 45 Athletic competence 1. 00 -. 13 Beck Depression Inventory 1. 00 (a) Correlation between Coopersmith Self-Esteem Inventories and Global SE composite: r = . 89. Correlation between Self-Esteem Index and Global SE composite: r = . 91. TABLE 3 Analysis of Variance(a) in Global and Domain Self-Esteem Scores as a Function of Gender: Results Presented as Standardized Mean Scores (z scores), F Scores, and p Values Legend for Chart: A - Measure B - Male M C - Female M D - Difference in means E - F A B C D E Global self-esteem Global self-esteem composite(b) . 14 -. 08 . 22 7. 5(**) Self-esteem domains Personal security . 22 . 02 . 20 6. 1(*) Home/Parents . 06 -. 11 . 17 8. 0(*) Peer popularity . 18 -. 03 . 21 2. 1 ns Attractiveness . 19 -. 05 . 24 9. 1(**) Academic competence -. 14 . 02 . 16 2. 4 ns Academic conscientiousness -. 11 . 06 . 17 9. 9(***) Personal mastery . 18 -. 07 . 25 28. 2(***) Psychological permeability -. 24 . 06 . 30 38. 6(***) Athletic competence . 25 -. 17 . 42 18. 2(***) Note. Means are presented in standardized form for ease of comparison between variables. (a) Derived from Gender x Grade two-way ANOVAs: Gender x Grade interaction effects did not reach significance for any variable. (b) The global self-esteem composite represents the combination of two separate inventories: Coopersmith Self-Esteem Inventories and the Self-Esteem Index. For clarity, their means and F values are as follows: Coopersmith: male = . 22, female = -. 08. F = 12. 8(***). Self-Esteem Index: male = . 07, female = -. 06, F = 2. 02 (ns). (*) p < . 01. (**) p < . 001. (***) p < . 0001. TABLE 4 Stepwise Regression Analyses(a) on Global Self-Esteem Composite and Depression Scores Legend for Chart: A - Measure B - Self-esteem composite beta C - Self-esteem composite DELTA R2 D - Self-esteem composite--Girls beta E - Self-esteem composite--Girls DELTA R2 F - Self-esteem composite--Boys beta G - Self-esteem composite--Boys DELTA R2 H - Beck Depression Inventory beta I - Beck Depression Inventory DELTA R2 A B C D E F G H I Home/Parents . 55(***) . 51 . 56(***) . 23 . 53(***) . 51 -3. 1(***) . 24 Personal security . 30(***) . 26 . 32(***) . 55 . 28(***) . 25 Personal mastery . 30(***) . 02 . 31(***) . 02 . 28(***) . 02 -1. 9(***) . 02 Academic competence . 23(***) . 07 . 24(***) . 08 . 22(***) . 06 -1. 0(***) . 01 Attractiveness . 22(***) . 04 . 20(***) . 03 . 27(***) . 05 -3. 5(***) . 18 Peer popularity 1. 2(***) . 01 Psychological permeability 2. 7(***) . 06 Gender(b) -1. 4(***) . 01 Note. DELTA R2 at time of entry. beta = weight in final model. Self-esteem composite, R2 = . 90(***). Self-esteem composite-girls, R2 = . 91(***). Self-esteem composite-boys, R2 = . 88(***). Beck Depression Inventory, R2 = . 51(***). (a) Eight self-esteem domains as well as gender were entered into the models. Academic conscientiousness was excluded because it sampled only a subset of the domain of interest. (b) Male = 1, female = 0. (***) p < . 0001. REFERENCES Allgood-Merton, B., Lewinsohn, P. M., & Hops, H. (1990). Sex differences and adolescent depression. Journal of Abnormal Psychology, 99(1), 55-63. Avison, W. R., & McAlpine, D. D. (1992). Gender differences in symptoms of depression among adolescents. Journal of Health and Social Behavior, 33, 77-96. Battle, J. (1980). Relationship between self-esteem and depression among high school students. Perceptual and Motor Skills, 51(1), 157-158. Baumrind, D. (1991). Effective parenting during the early adolescent transition. In P. A. Cowan & M. Hetherington (Eds.), Family transitions (pp. 111-162). Hillsdale, NJ: Erlbaum. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. New York: Guilford Press. Beck, A. T., & Steer, R. A. (1987). Beck Depression Inventory manual. San Antonio, TX: The Psychological Corporation. Beck, A. T., Steer, R. A., & Garbin, M. (1988). Psychological properties of the BDI: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77100. Beyer, S., & Bowden, E. M. (1997). Gender differences in self-perceptions: Convergent evidence from three measures of accuracy and bias. Personality and Social Psychology Bulletin, 23(2), 157-172. Bolognini, M., Plancherel, B., Bettschart, W., & Halfon, O. (1996). Self-esteem and mental health in early adolescence: Development and gender differences. Journal of Adolescence, 19(3), 233-245. Brack, C. J., Orr, D. P., & Ingersoll, G. (1988). Pubertal maturation and adolescent self-esteem. Journal of Adolescent Health Care, 9(4), 280-285. Brage, D., & Meredith, W. (1994). A causal model of adolescent depression. The Journal of Psychology, 128(4), 455-468. Burnett, P. C. (1996). Gender and grade differences in elementary school children's descriptive and evaluative self-statements and self-esteem. School Psychology International, 17(2), 159-170. Cairns, E., McWhirter, L., Duffy, U., & Barry, R. (1990). The stability of self-concept in late adolescence: Gender and situational effects. Personality and Individual Differences, 11(9), 937-944. Chodorow, N. (1978). The reproduction of mothering: Psychoanalysis and the sociology of gender. Los Angeles: University of California Press. Chubb, N. H., Fertman, C. I., & Ross, J. L. (1997). Adolescent self-esteem and locus of control: A longitudinal study of gender and age differences. Adolescence, 32(125), 113-129. Cohen, J. (1977). Statistical power analysis for the behavioral sciences. New York: Academic Press. Consulting Psychologists Press. (1981). Self-Esteem inventories. Palo Alto, CA: Consulting Psychologists Press. Coopersmith, S. (1967). The antecedents of self-esteem (pp. 3-10). Palo Alto, CA: Consulting Psychologists Press. Demo, D. H., & Savin-Williams, R. C. (1983). Early adolescent self-esteem as a function of social class: Rosenberg and Pearlin revisited. American Journal of Sociology, 88, 763-774. Demo, D. H., Small, S. A., & Savin-Williams, R. C. (1987). Family relations and the self-esteem of adolescents and their parents. Journal of Marriage and the Family, 49, 705-715. DeMoss, K., Milich, R., & DeMers, S. (1993). Gender, creativity, depression and attributional style in adolescents with high academic ability. Journal of Abnormal Child Psychology, 21(4), 455-467. Dweck, C. S., & Bush, E. S. (1976). Sex differences in learned helplessness: I. Differential debilitation with peer and adult evaluators. Developmental Psychology, 12, 147-156. Eccles, J. S., Midgley, C., Wigfield, A., Buchanan, C. M., Reuman, D., Flanagan, C., & MacIver, D. (1993). Development in adolescence: The impact of stage-environment fit on young adolescents' experiences in schools and in families. American Psychologist, 48(2), 90-101. Frey, K. S., & Ruble, D. N. (1987). What children say about classroom performance: Sex and grade differences in perceived competence. Child Development, 58, 1066-1078. Gjerde, P. F., Block, J., & Block, J. H. (1988). Depressive symptoms and personality during late adolescence: Gender differences in the externalization-internalization of symptom expression. Journal of Abnormal Psychology, 97(4), 475-486. Harter, S. (1982). The Perceived Competence Scale for Children. Child Development, 53, 87-97. Harter, S. (1990a). Causes, correlates, and the functional role of global self-worth: A life-span perspective. In R. J. Sternberg & J. Kolligan, Jr. (Eds.), Competence considered (pp. 67-97). New Haven, CT: Yale University Press. Harter, S. (1990b). Self and identity development. In S. S. Feldman & G. R. Elliott (Eds.), At the threshold: The developing adolescent (pp. 352-387). Cambridge, MA: Harvard University Press. Harter, S., & Whitesell, N. R. (1996). Multiple pathways to self-reported depression and psychological adjustment among adolescents. Development and Psychopathology, 8, 761-777. Hill, J. P., & Lynch, M. E. (1983). The intensification of gender-related role expectatio