Running head: EATING DISORDERS Eating Disorders and Perfectionism: Dependence and Relations’ Analysis Completed by: YOUR NAME University of Outline 1. Abstract 2. Introduction a. Eating Disorders normal and pathological path analysis b. Limitations of the study 3. Perfectionism as one of the causes of eating disorders a. Triggers b. Medical factors analysis 4. Perfectionism’s relations to eating disorder examined 5. Causes and dependence 6. Conclusion 7. References Abstract This paper attempts to determine a dependence and close relations between eating disorders and perfectionism as illustrated in the contemporary literature, studies conducted by medical professionals, and various researches that examine the issue of inter-dependence and close relations between these two concepts. It is important to note that eating disorders most commonly begin sometime during adolescence, rather than in childhood or adulthood. Bulimia nervosa (BN) is virtually unheard of prior to adolescence (Gislason 1988) and the vast majority of women clinically diagnosed with BN have symptom onset before age 25 (Woodside & Garfinkel 1992). Similarly, in clinical samples the modal age of onset of binge eating is 18; it is rarely seen in children (Streigel-Moore 1993). Anorexia nervosa (AN) does occur in prepubertal children but it increases dramatically after puberty, with the majority of cases beginning before age 25 (Woodside & Garfinkel 1992). Introduction Early adolescence and late adolescence may constitute special risk periods for the development of eating disorders (Wooley & Wooley 1985). This suggests that adaptation to the events or developmental changes commonly associated with these transitions can take the form of eating disorders. The increase in associated eating problems, such as body dissatisfaction and dieting during early adolescence, bolsters this impression (Koff & Rierdan 1991). The focus of this paper is on the early and late adolescence transitions. The study of developmental transitions represents an important component of the developmental psychopathology approach (Smolak & Levine, 1994a). Indeed, the way in which adolescents organize experiences during these transitions may be especially important in predicting future adaptation. Developmental transitions may be viewed as times of special risk because they involve changes in molar (core or underlying) structures and in normative challenges and stressors (in addition to any non-normative stressors that might appear). In a sense, they represent "crossroads" (Brown & Gilligan, 1992) of development, points at which a variety of developmental paths might be followed. This paper is divided into four major sections. In the first we explore the meaning of eating disorders. In the second we examine the perfectionism as applied to eating disorders, and in the third we discuss the late adolescent transition, predispositions, and factors that trigger the development of eating disorders in adolescents and older adults alike. These transitions are considered separately because they may differ in their relations to BN and to AN(Smolak & Levine, 1993). Furthermore, age of onset may provide clues to the etiology of these disorders (Woodside & Garfinkel, 1992). In the last section we consider the implications of developmental transitions for prevention, research, and theory concerning eating problems as connected and dependable on perfectionism cases severity. Eating Disorders normal and pathological path analysis The principle of a relationship between normal and pathological development of eating disorders also fits with the argument that there is at least a partial continuum between normative attitudes and behaviors and eating disorders. Many women diet but few become bulimic or anorexic. There is even a group of women and girls who have sub-threshold eating disorders but who will, as best as is known, never develop full-blown AN or BN (Striegel-Moore & Marcus, 1995). What protects these latter groups of women? The developmental literature offers a variety of specific possibilities. For example, high public self-consciousness, which is related both to depressive affect and to eating disorders (Striegel-Moore et al., 1993), is associated with high parental emphasis on achievement. Thus, parents who are supportive rather than pressuring may have children who can face the increased achievement demands of adolescence (Smolak & Levine, 1994a) more adequately. This is generally consistent with findings that continued close (but not too close) relationships with parents might benefit adolescents. Whether certain types of social support are more protective than others remains an issue for future research. Limitations of the Study Etiological models of the eating disorders have been predominantly based on the experiences and recollections of women seen in clinical practices. Such models have made crucial contributions to the understanding of eating disorders, but they are limited by both their referent sample and their theoretical perspectives. Developmental psychopathology offers a new approach to studying eating disorders, encouraging greater emphasis on developmental process, protective factors, and childhood precursors. As its name implies, developmental psychopathology has its roots in developmental psychology. As a discipline that is heavily (although by no means exclusively) focused on children and adolescents, developmental psychology has much to offer in terms of theory and research techniques. As the parts in this paper will demonstrate, there is a wealth of information on a variety of characteristics and experiences that have long been assumed to play some role in the development of eating problems. This information can be helpful -- and, indeed, may be necessary -- in formulating more specific hypotheses concerning these factors and in designing research to assess their contributions to eating problems. Developmental psychopathology also holds the potential to facilitate primary prevention. By understanding childhood precursors, it prepares us to design prevention programs that focus on ameliorating weaknesses and encouraging strengths. It also provides information about the concerns of children and adolescents as well as about their ability to comprehend certain material. This can help us gear our programs to the interests and levels of our audience. Working with children and adolescents adds new demands to research. One such demand is the development of new measures. Several of the developmental approaches illustrated in this paper contain discussions of specific measures of children's development. In addition, such measures and techniques may serve as prototypes for the development of more psychometrically sound and theoretically meaningful measures of eating attitudes and behaviors for use with children. The information in the developmental parts can also be helpful in addressing major change in developmental structures. One of the great challenges of prospective research is to find measures that adequately represent qualitatively different forms of a behavior (Smolak & Levine, 1994a). Consider, for example, the concept of the "superwoman." Several theorists have suggested that this characteristic (or role) is a risk factor for the development of eating problems (Levine & Smolak, 1992). As typically defined, however, the "superwoman" is an adolescent or an adult. What does she look like in childhood? What elements are critical to the development of this characteristic and identifiable during the developmental period? Possibilities include social approval, multiple roles, achievement orientation, low self-esteem coupled with perfectionism, or some early combination of these (Proffitt & Smolak, 1995). Developmental research on self, gender roles, and achievement can be used to help frame these questions and design the studies that can address them. Linking qualitatively different structures across periods of development can be difficult. Even if a group is initially relatively homogeneous for a characteristic, not all of them necessarily follow the same developmental paths to identical outcomes. For example, some securely attached infants develop psychopathologies whereas some insecurely attached infants do not (Smolak & Levine, 1994a).
Thus, the task facing psychologists is to chart not simply the relationship between two developmentally distinct states but the evolving pathways from one or more states to their particular outcomes. This endeavor requires not only longitudinal data but special analytic methods. Most important, it requires theory that can in many cases already be found in the developmental literature. Perfectionism as one of the causes of eating disorders Even in elementary school, some of the behaviors, attitudes, and beliefs that may lead a girl to emerge from the early adolescence transition with eating problems may be in place. These include personality predispositions such as perfectionism and low self-esteem (Brown & Gilligan, 1992). Girls' strong investment in social relationships may also render them vulnerable as they move into a setting (middle school) where such relationships change dramatically (Brown & Gilligan, 1992). Attitudes about the importance of thinness and about weight management behavior may also be risk factors (Smolak & Levine, 1994b). Although we cannot examine all of these potential predispositions in detail, some discussion is possible. Perfectionism has long been linked to adult eating problems (Garner, Olmstead, & Polivy, 1983). Indeed, the original image of the anorexic was that of "the best little girl in the world" (Levenkron, 1978, p. 34). Perfectionism may be linked to eating problems through the "superwoman" role (Levine & Smolak, 1992). Brown and Gilligan (1992) documented the existence of perfectionism among some elementary school girls. These girls want to always be pretty, be nice, do well in school, and please others. Such unrealistic goals may set the stage for a loss of self-esteem as it becomes clear to the girls that they are falling short of their own ideal. These goals may also be driven by a need for social approval, which has been associated with eating problems in adults. This pattern may be especially insidious for girls who adopt White middle-class ideals when their race or ethnicity guarantees that they will never find full acceptance, even if they do become thin, earn good grades, and so forth (Brown & Gilligan, 1992; Thompson, 1994). In terms of behavioral perfections, research has established repeated dieting, especially with weight fluctuations, as a risk factor for AN, BN, and their sub-threshold variants. Early onset of dieting behavior is associated with increased risk (Tobin, Johnson, Steinberg, Staats, & Dennis, 1991). Dieting children and adolescents tend to be heavier than non-dieters, but they are usually not obese or even overweight by medical standards (Smolak & Levine, 1994b). The risk created by dieting, then, is not solely attributable to real weight problems. It is possible that the risk is due to dieting itself, which can lead to metabolic slowing requiring more extreme methods to lose weight and maintain weight loss (Garner & Wooley, 1991). As the body struggles to escape starvation, binge eating may develop (Striegel-Moore, 1993). Given this possible path toward the development of eating problems, it is worrisome that up to 40% of elementary school girls report trying to lose weight (Smolak & Levine, 1994b). In addition, research suggests that young children know that it is undesirable to be fat (Yuker & Allison, 1994). Not all elementary school children believe that thinness is important to attractiveness (Smolak & Levine, 1994b), but those girls who do may be especially susceptible to the peer and media pressure that is part of the early adolescent world, and hence to the development of eating problems (Levine, Smolak, & Hayden, 1994). Triggers As is true at other developmental transitions, there are many possible triggers of the early adolescence perfectionism. There are changes in peer and family relationships, academic demands, and social role expectations (e.g., gender role intensification) that render childhood cognitive, personality, and social structures inadequate. All of these changes might be viewed as marking the gradual ending of childhood. Puberty is probably the most salient indicator, both to the child and to others. Thus, puberty itself requires adaptation by both the girl and her social networks. Among the events that the girl must accept are substantial weight and fat gain (relative to muscle mass), reproductive maturity, and an adult body that may evoke increases in sexual harassment. Changes in the social network include shifts in the parent-child relationship. For example, mother-daughter conflict as well as daughters' resistance to paternal conversational interruptions increase (Hill, 1988). Such changes may make the daughter feel distant from her family, especially if, as is true for many individuals with AN, conflict and disobedience are unacceptable in the family. It is noteworthy that the hormonal changes associated with puberty have few substantial direct effects on behavior. When hormonal effects are documented, they tend to be either fairly small (Brooks-Gunn & Warren, 1989) or mediated by other factors. Furthermore, puberty's effects are not the same in all cultures (Rodriguez-Tome, Bariaud, Zardi, Delmas, & Szlagyi, 1993). This implies that many of the behavioral effects associated with puberty are attributable to the social construction of puberty rather than to the biological processes per se. Puberty is a bio-psychosocial event, not just a biological shift. The social meaning of puberty in the United States has several implications for the development of eating disorders. More than anything else, puberty indicates to others than a child is becoming an adult. In general, achieving adult female status is at best an ambivalent event, whereas the attainment of adult male status is generally positive. Thus, as girls go through puberty they find themselves moving away from the ideal body shape for women; their options for careers, sports, and other types of opportunities becoming more rather than less restricted; and the value of their strengths (e.g., in relationships) diminishing (Brown & Gilligan, 1992). One thing becomes increasingly clear to girls as they move into the early adolescence transition: Attractiveness is an important component of female success. This is true academically, socially, and in terms of career and some athletics (Yuker & Allison, 1994). Small wonder, then, that as girls go through puberty they work harder and harder at their appearance. Given that puberty means that most girls no longer have the long-legged, slender, lean prepubescent body that White middleclass society deems beautiful, it should not be surprising that eating problems begin to increase substantially at puberty. Perfectionism’s relations to eating disorder examined Findings relating puberty to eating problems are summarized in Table 1. It is clear that as White American girls go through puberty they become more dissatisfied with their weight and, to a lesser extent, with their shape (they are generally pleased about breast development). They are also more likely than same-aged pre-pubertal girls to diet and to worry about their weight. Findings about the timing of puberty are more ambiguous. Girls who are "off time," especially those who mature early, are more dissatisfied with their bodies, partially because early maturers weigh more than their pre-pubertal peers (see Table 1). Whether the greater weight or dissatisfaction of early maturers persists once all girls have completed puberty is not clear. Transitions present multiple challenges in the form of developmental tasks. Indeed, variants of the tasks occur in most cultures because of, for example, the universal expectation that adults contribute to society in terms of work and procreation. How might something so positive contribute to the development of eating problems? The eating problems represent an adaptation to the transition's demands. This idea is central to the developmental psychopathology model (Smolak & Levine, 1994a) as well as to many feminist models of eating problems (Striegel-Moore, 1994). A girl who decides to diet may not be making an unreasonable decision, even if that dieting eventually endangers her life and well-being. Different paths may lead to problematic adaptation. For some girls, predispositions are critical in their appraisal of the early adolescent tasks. Girls who are very invested in thinness, for example, may be appalled at the fat that maturation brings.
TABLE 1 Studies Relating Puberty to Eating Problems/Disorders Author, Date Sample Puberty Measure Findings Alsaker, 1992 1,109 girls and 1,256 boys Students rated a measure of timing of No relationship between CPT and body in Norway in Grades 6 global pubertal development (computed dissatisfaction for the girls. Among 8th-grade through 9 (cross-sectional) pubertal timing; CPT). Also rated their boys, early maturers were more satisfied with relative timing to their same-sex their bodies. PPT was related to body classmates (Perceived Pubertal Timing; dissatisfaction among 6th-grade girls; among 8th- PPT). grade girls, the girls who perceived themselves as on-time showed greater body dissatisfaction. Among 7th- and 8th-grade boys, self-perceived late maturers had poorer body image. Atkins & Silber, 21 female childhood Tanner Stages (breast & pubic hair); For 10 of the girls, onset of AN coincided with 1993 anorexics (9-12 years at menarcheal status early, but normal, puberty diagnosis) Attie & Brooks- 193 7th through 10th Classified as early, late, or on-time In the longitudinal analysis, actual timing was Gunn, 1989 grade, private school, maturers based on time of menarche, not related to EAT-26 scores. In the concurrent White girls and their Tanner breast stage, and Tanner pubic analyses, the combination of actual and perceived mothers. Retested 2 years hair stage at Time 1. Girls were also timing and body fat was related to EAT scores at later. asked about their own perceptions of their Time 1 and Time 2. pubertal timing. Blyth, Simmons, & 210 White girls, tested in Self-report time of menarches Earlier developers were more dissatisfied with Zakin, 1985 6th, retested in 7th grades their weight in 7th grade. Brooks-Gunn, Attie, Study 1: Timing based on self-reported menarcheal Late-maturing dancers had more positive body Burrow, Rosso, & 424 girls aged 14-18, 287 age. image and dieted less than on-time maturing Warren, 1989 nonathletes, 72 dancers. competitive swimmers, 64 students from national ballet schools. Almost all White and high SES. Study 2: Tanner Stage based on self-report Physical development (pubertal stage + 238 girls, mostly White, menarche and Tanner Breast weight/height) was significantly related to EAT-26 in Grades 7-10 at Time 1 Development Scale. scores for the nonathletes only in the Time 1 testing. Retested 2 years Concurrent Analysis. In the longitudinal analysis, later. 193 were physical development (at Time 1) was not related nonathletes; 45 were to Time 2 EAT-26 scores. dancers. Author, Date Sample Puberty Measure Findings Brooks-Gunn & 276 private school girls Menarcheal age. On-time dancers were heavier, had higher Warren, 1985 and 69 dance school girls, pathology, BN, and higher perfectionism scores aged 14-17. and poorer body image than late-maturing dancers. More dancers were late maturers. Cattarin & 210 10-15-year-old girls, Pubertal Development Scale; Age at Time 1 maturational status did not predict Time 2 Thompson, 1994 87 of whom were retested menarche teasing, body image, eating disturbance, or global after 3 years. psychological functioning. Crockett & Pe- Longitudinal sample (253 Pubertal Development Scale In Grades 7 and 8, more mature girls were less Petersen, 1987 boys & girls) beginning in satisfied with their weight. 6th grade and ending in 8th Dom, Crockett, & 253 girls and boys The Pubertal Development Scale (PDS; In the 7th and 8th grades, girls who were more Petersen, 1988 followed from 6th through Petersen, Crockett, Richards, & Boxer, advanced in pubertal development were more 8th grade. No information 1988). dissatisfied with their weight. No multivariate on race. effects for boys. Dornbusch et al., 6,768 12-17-year-olds A composite based on Tanner breast and Within social class, as maturity level increased for 1984 (45% female), examined pubic hair development scales. girls, the desire to be thinner increased. The as part of the U.S. majority of girls who were postpubertal wanted to National Health be thinner. Examination Survey between 1966 and 1970. Fabian & 61 pre- and 60 Self-reported menarcheal history. Postmenarcheal girls were both heavier and Thompson, 1989 postmenarcheal girls, higher or, Drive for Thinness than age-matched aged 10-15 Author, Date Sample Puberty Measure Findings Graber, Brooks- 116 girls tested initially Self-reported age at menarche. Early maturers were more likely, as young adults, Gunn, Paikoff, & at 7th, 8th, or 9th grade to be "chronically" at risk for eating problems Warren, 1994 and then again in 9th, (EAT-26 score %u22DA 20 at all three times of testing). 10th, or 11th grade and However, this effect appears to be due to higher then again between ages body fat among the early maturers rather than to 21 and 23. Private school, early puberty per se. predominantly White Gralen, Levine, 121 6th-, 116 8th-, 144 Self-reported menarcheal status (yes/no). Menarcheal status was related to dieting and EAT- Smolak, & Murnen, 9th-, and 73 10th-grade Used only in 6th- and 8th-grade factor-2 (Bulimia) scores for the 6th graders only. 1990, girls, almost all White regressions. Gross & Duke, 1980 National Health Tanner Scales Early-maturing girls were more dissatisfied with Examination Survey of their weight. 3, 196 girls aged 12-17 Killen et al., 1992 939 6th- and 7th-grade Self-ratings of breast and pubic hair More advanced pubertal development was girls. 49.6% White; 22.8% development (Tanner Stages) summarized associated with more symptoms of eating Latina; 19.7% Asian; 3.2% as one measure: The Sexual Maturity problems/disorders. Pacific-Islander; 3.6% Index (SMI). Black; 2.6% Native American; 7.5% other. Killen, Hayward, et 839 6th- and 7th-grade Tanner Breast and Pubic Hair Girls with bulimic symptoms were more al., 1994 girls Developmental Scales developmentally mature than asymptomatic girls. Killen, Taylor, et al., 3-year study of 939 6th- & Sexual Maturity Index (SMI) based on SMI was not related to the likelihood of 1994 7th grade girls Tanner Breast and Pubic Hair developing symptoms over the 3-year period. Development Stages. Koff, 1993 146 girls followed from Early (menarche before Fall of Grade 7) Early puberty was associated with higher body 6th through 9th grade vs. nonearly maturing based on self- weight and greater BMI. The early maturers had reports of age at menarche. higher scores on EDI Subscales for Drive for The fat gain may be especially difficult for the girl who has been restraining her eating to control her weight during childhood and who now finds her efforts inadequate. Girls' appraisal of the developmental tasks is also affected by their socio-cultural context. Teasing about weight, for example, increases the risk of eating problems (Fabian & Thompson, 1989; Levine, Smolak, Moodey, et al., 1994). Some middle school girls live in a "culture of dieting" in which their mothers and friends diet and worry about weight and the girls read the teen magazines that send the broader culture's slenderness message. Such girls tend to have higher Eating Attributes Test scores, to diet more frequently, and to express greater body dissatisfaction (Levine, Smolak, & Hayden, 1994). These girls may be developing a thinness schema, similar to those hypothesized to occur in restrained eaters and eating-disordered patients. This schema includes information about the importance of thinness to attractiveness, the importance of attractiveness to social and career success, the dangers of fat, methods of becoming thin, and the relevance of attractiveness to one's own self-esteem. The thinness schema is hypothesized to lead to selective processing of information that supports the girl's assessment of thinness and its importance. Predispositions may affect a girl's appraisal of, and hence adaptation to, developmental tasks. Also, not all girls face the same types of tasks.
Based on the work of Simmons and Blyth (1987), we have proposed that cumulative normative developmental stressors increase the likelihood of eating problems. Specifically, disordered eating attitudes and behaviors may be fostered when simultaneous weight gains, changes in female-male relationships, and threats to achievement are filtered through the thinness schema (Smolak & Levine, 1994a). Events are considered to be simultaneous when they occur within the same year. If either weight gains or hetero-social changes occur alone, the outcome is likely to be an intensification of weight and shape concern and the development of non-pathological dieting. In early adolescence, weight gains can be related to pubertal development and hetero-social changes to the onset of dating, both of which have been shown to contribute to increased body dissatisfaction and dieting (Gralen et al., 1990; Levine & Smolak, 1992; Levine, Smolak, Moodey, et al., 1994). If these challenges occur simultaneously and the thinness schema is severe, the risk of sub-threshold eating disturbances rises. If there are also threats to one's sense of achievement, such as the new demands of middle school (Eccles & Midgley, 1990), and if the girls' self-definition emphasizes success in multiple roles in order to gain external approval (the "superwoman complex"; Steiner-Adair, 1986), the risk of frank eating disorders increases. Without the thinness ideal or the superwoman complex, the simultaneous stressors are likely to result in a different disorder, adjustment problems, or successful coping. We have illustrated evidence suggesting that simultaneity of the normative tasks is more influential than the relative timing (early vs. late) in the development of eating problems (Smolak, Levine, & Gralen, 1993). Not surprisingly, girls who have been both early and simultaneous in the onset of dating and puberty (and who concurrently enter middle school) are particularly likely to develop problematic attitudes and behaviors. These girls experience the dual stress of being "off time" (Smolak, 1993) and of having to cope with multiple changes. Causes and dependence Although clinical cases of eating disorders are relatively rare in pre-pubertal children, various eating disturbances occur among children that may lead to subsequent development of eating disorders during adolescence ( Marchi & Cohen, 1990). Among the risk factors for eating disorders that have been identified in children are early feeding problems, family dysfunction, perfectionism, low self-esteem, and being teased about being fat (Gowers et al., 1991). Specific strategies to help prevent the development of eating disorders or risk factors for later eating disorders in pre-pubertal children have been suggested by various investigators, including Levine ( 1994), Smolak and Levine (1994a, 1994b), and Striegel and Moore ( 1992). Some of the strategies that have been suggested for parents and caregivers include educating expectant parents about children's eating patterns and preferences at different stages of development and emphasizing that how one responds to an infant's hunger cries is as important as what the infant is fed; teaching day-care workers to detect feeding problems in children so that they can serve as a support and referral system for early intervention; making parents aware of how their own eating behaviors and weight concerns can affect their children; educating parents to refrain from commenting on children's weight and body shape and to focus instead on health, strength, or abilities; and alerting parents and family therapists to the possibility that a child may develop eating problems in response to family conflict or to loss of a parent through separation, divorce, or death. Suggested strategies for children include teaching them acceptance of a wide range of body shapes; encouraging them to develop interests and skills that will lead to success and personal fulfillment without being based on appearance; teaching them healthy eating and exercise habits and emphasizing the negative effects of dieting; teaching ways to resist teasing about weight and being pressured to diet; and providing children with information about the facts of development to allay some of their anxieties about puberty, emphasizing that weight gain is a normal and necessary part of pubertal development. Conclusion Researchers and theorists have long believed that puberty, especially early puberty, plays a special role in the development of eating disorders, particularly of anorexia nervosa. Research does confirm that specific forms of body dissatisfaction and dieting do indeed increase as girls go through puberty (Smolak & Levine, 1994b). Thus, girls who are early maturers begin dieting and engaging in other weight management behaviors sooner than their later developing peers. What is less clear is whether these early maturing girls are permanently at greater risk for eating problems. The data linking maturational timing and later eating problems are mixed. More recent evidence suggests that the relationship may be indirect, mediated, for example, by adiposity (Graber et al., 1995), teasing, or co-occurrence with normative stressors (Smolak et al., 1993). Such findings are a reminder of the flexible and transactional nature of development. They also serve to caution against a long-standing tendency to pathologize normal female reproductive development and functioning. Effects of the Media Many highly regarded experts on eating disorders believe that exposure to media images of slenderness constitutes one important risk factor for the spectrum of eating disorders and perfectionism development. The conviction that these images do much more than passively reflect dominant cultural values is seen in the fact that, as noted above, most curricula for prevention of eating disorders include lessons on analysis of and resistance to media messages. However there is a great deal yet to be learned about the ways in which the multitude of media images of slenderness and accompanying articles on diet and perfectionism’s influence girls and women. For example, exposure to slender models appears to have little immediate effect on the body images of adolescent girls or college-aged women unless these individuals already have significant weight and shape concerns or frank eating disorders. With respect to the cumulative effect of media, a study by Stice, Schupak-Neuberg, Shaw, and Stein (1994) offers some preliminary evidence for a direct connection between amount of media exposure during the previous month and scores on the Eating Attitudes Test (EAT). Other studies reviewed by Levine and Smolak (1994a) point to the importance of the transaction between an adolescent girl's motives for using media and the nature of media selected. In a study completed, Meade and Levine (1995) administered media use surveys, a measure of weightist prejudice, and the child version of the Eating Attitudes Test to 355 middle school girls. The vast majority of the sample were White and from working-class and middle-class backgrounds. The multiple regression analysis (which accounted for 34.5% of the variation) yielded no significant main effect for amount of media use (or magazine use). Instead, this analysis pointed to the importance of weightist prejudice and motives for self-evaluation and self-improvement as significant predictors of disordered eating in middle school girls. Replication and extension of this research is certainly necessary before firm conclusions are reached. Nevertheless, it appears that (a) exposure to media messages about the multidimensional importance of slenderness is indeed very likely a risk factor for disordered eating; and (b) the impact of this exposure should be conceptualized within the transactional perspective emphasized by developmental psychopathology. In this regard the work of two academic researchers in the area of marketing, Martin and Kennedy ( 1994a, 1994b), is highly recommended for its attention to the role of social comparison processes in the effects of media on young girls. From the transactional perspective, the context of that media use also appears to be very important. For example, Levine, Smolak, and Hayden ( 1994) found very high ChEAT scores in a subgroup of White middle school girls who reported both that magazine articles and advertisements greatly influenced their conception of the ideal body shape and how to get it, and that their peers and their parents were also highly "invested in" thinness.
As developmental psychology captures the fancy of more and more students and professionals interested in eating disorders, the following statement is sometimes heard: "Longitudinal designs will solve our problems in regard to demonstrating which phenomena cause eating disorders." Smolak (1995) offers several cautions, however. First, longitudinal designs have limitations in regard to sampling and measurement. Second, various types of designs and statistical analyses are probably necessary to capture the blend of general influences, individuality, unidirectional causality, and reciprocality postulated by a developmental perspective. Third, statistical wizardry (e.g., in structural equation modeling) cannot substitute for elegance of theory and for clear relationships between data and a priori hypotheses. According to Smolak (1995), methodological sophistication, a renewed interest in psychometry, a commitment to flexible theorizing, a multidisciplinary perspective, and a great deal of patience are core requirements for the "developmental psychopathology major." Nevertheless, advances in the understanding of depression (have clearly demonstrated the exciting potential of the developmental perspective. Moreover, we believe that there is plenty of room for undergraduates, graduate students, and professionals alike to respond to industry challenge and to begin to fill in the many gaps in the current state of knowledge about the developmental psychopathology of eating disorders.