Sources of Disordered Eating Patterns between Ballet Dancers and Non-Dancers
Anshel, Mark H., Journal of Sport Behavior
Disordered eating may be defined on a continuum from eating disorders (e.g., anorexia nervosa, bulimia nervosa) to preoccupations with weight and restrictive eating (Wein & Micheli, 2002). However, there are several differences between disordered eating and eating disorders. For example, while a disordered eating pattern is a habitual reaction to life situations, an eating disorder is a mental illness (American Psychiatric Association, Diagnostic and Statistical Manual, 4th ed., 1987). In addition, disordered eating is not usually accompanied by frequent thoughts of food, eating, and one's physical appearance. Persons with eating disorders, on the other hand, have compulsive thoughts of food, eating, and one's body. While disordered eating may lead to transient weight changes and nutritional problems, major medical complications are very rare. Eating disorders, by contrast, often lead to serious medical problems, with a mortality rate of 2-10% (Comer, 2001). Finally, each condition is treated differently. Disordered eating requires education and the problem may diminish without treatment. Eating disorders, however, require specific medical and mental health treatment, without which the problem will persist. These differences are relevant because there exists far more published research exists examining eating disorders than disordered eating among ballet participants. This study addressed both eating disorders and disordered eating patterns, including their relationship to selected dispositions.
Several researchers (e.g., Garner & Garfinkel, 1980; Hamilton, Brooks-Gunn, & Warren, 1985) contend that the incidence of disordered eating patterns is estimated to be higher in models, dancers, and athletes than in the general female population. For example, the prevalence of anorexia nervosa is higher for ballerinas compared with the general American population of adolescent females (Brooks-Gunn, Warren, & Hamilton, 1987). In fact, dance students are seven times more likely than high school students to develop anorexia nervosa (Clough & Wilson, 1993). Brooks-Gunn et al. (1987) reported that one-third of a sample of professional dancers had a disordered eating pattern, while Hergenroeder, Wong, Fiorotto, Smith, and Klish (1991) found that 43% of the Houston Ballet Academy was diagnosed with anorexia nervosa. By contrast, Holderness, Brooks-Gunn, and Warren (1994) reported that dancers and nondancers did not differ significantly with regard to the diagnosis of anorexia nervosa and bulimia nervosa.
While the physical and technical demands of dance may be similar to other highly competitive sports, the aesthetic requirements of body image beyond the functional requirements for dance are fundamentally different (Calabrese, Kirkendall, Floyd, Rapoport, Williams, Weiker, & Bergfeld, 1983). According to Hamilton, Brooks-Gunn, and Warren (1986) and Pierce, Daleng, and McGowan (1993), dancers who fail to meet and maintain a predetermined ideal body composition are rapidly "deselected" from professional participation. Further, it appears that age 21 yrs is the point at which it is determined whether a dancer will become "successful," after which time these chances are greatly reduced (Druss & Silverman, 1979).
Another factor that exacerbates the likelihood of disordered eating patterns among dancers is that exercise alone, in the absence of strict dieting, may not result in the desired weight loss. For example, Cohen, Segal, Witriol, and McArdle (1982), in their study of 15 professional ballet dancers from the American Ballet Theatre, found the caloric expenditure for an entire one hour ballet class to average only 200 kcal/hr for women, compared to the caloric expenditure of about 500/hr kcal for swimming and skating (Brooks-Gunn et al., 1987). According to Cohen et al. (1982) and Cohen, Potosnak, Frank, and Baker (1985), these findings provide evidence to suggest that classical ballet is a relatively inefficient method of burning calories, and, for this reason, dancing, without a dietary intervention, cannot produce weight reduction and maintain the low weights required for the classical physique. The chronic dieting behavior of dancers may be important in the pathogenesis of eating disorders (Garner & Garfinkel, 1980; Hamilton et al., 1985). The demands for thinness and the low caloric expenditure of dance make eating disorders almost normative in the dance world (Brooks-Gunn, Burrow, & Warren, 1988).
While it is not assumed that biological or familial factors are irrelevant in the development of eating disorders, it is apparent that the dance environment is thought to represent a subculture that amplifies sociocultural pressures to conform to predetermined and stereotypical expectations by the dancer's significant others, and places them at risk for eating disorders. This is likely due to their preoccupation with weight control (Hausenblas & Carron, 1999). Given the medical dangers of these conditions, understanding the factors that may predispose dancers to this illness in the hope to prevent or treat it appears warranted.
The onset of eating disorders generally occurs in adolescence or young adulthood, and is largely confined to women (Fay & Price, 1993; O'Dea, 2002). More than 90% of anorexia nervosa and bulimia nervosa sufferers are adolescent or young adult females, and approximately 5% of women suffer from a partial or subclinical form of either anorexia nervosa or bulimia nervosa (Slavin, 1987; Woodside & Garfinkel, 1989). Females in western cultures are at much greater risk for developing eating disorders than males due to intense social pressures to conform to the current cultural ideals of feminine beauty (O'Dea, 2002). Body shape and weight become critical determinants of self-esteem, particularly in adolescence, because interpersonal success is increasingly seen as closely linked to physical attractiveness (Brownell, Rodin, & Wilmore, 1992). According to Johnson, Steinberg, and Lewis (1988), thinness has increasingly been associated with a highly valued personal achievement, demonstrating self-control, autonomy, and success. Thus, the pursuit of thinness is commonly perceived an action or goal in which young women can obtain favorable social responses thereby enhancing self-esteem.
Although there is a paucity of research on dispositional predictors of eating disorders in ballet dancers, in one older study, Druss and Silverman (1979) found that classical ballet students were at risk for developing eating disorders due to their extreme dedication to dancing, despite their knowledge that over 95% of dancers do not reach ballerina status. The students also revealed an acute sensitivity to criticism, an intense drive toward achieving personal goals, and have a need for high personal control over eating and other sources of pleasure. The researchers surmised that dancers were motivated by a desire to achieve perfection because "they practice exercises ritualistically to stop thinking, feeling, to reduce their bodies to the barest minimum of bone and muscle and to deny themselves all other pleasures in order to achieve that goal. They dare to hope for some brief time to become perfect" (p. 12 l). For example, Clough and Wilson (1993), in their study of Australian ballet dancers, found a high, significant correlation (r = .69) between eating disorder characteristics and the average total score on the Multidimensional Perfectionism Scale (Hewitt & Flett, 1991).
According to Bruch (1973, 1978), struggling to live up to perfectionistic or unrealistic standards is characteristic of the eating disorder, anorexia nervosa. The combination of sociocultural pressures and pressures inherent in the dance profession itself identify dancers as a vulnerable subgroup of women who are at high risk of eating disorders. The pressure for thinness, combined with expectations of high performance, such as dance, produce the ideal social climate for the manifestation of eating disorders, particularly in vulnerable adolescents (Garner& Garfinkel, 1980). Whether these pressures lead to emotional disturbance and eating disorders is uncertain. Maloney (1983) found that dance students are seven times more likely to develop an eating disorder than high school students, in general. However, in a study that contradicted these results, Lancelot, Brooks-Gunn, Warren, and Newman (1991) found a high association between eating disorders and emotional disturbance in non-dance, but not in dance, environments. The researchers concluded that eating pathology is adaptive in a dance environment. This area requires further study.
There is an apparent paucity of research examining the dieting and weight concerns of ballet dancers and the presence or absence of psychological disturbance, particularly among Australian dancers. In addition, the extent to which psychological disturbance (e.g., neurotic perfectionism, fear of failure, low self-esteem) accompanies dietary and weight concerns, and the psychological characteristics of dancers who experience an eating disorder as compared to non-dancers of similar gender and age has been rarely examined. Further research on psycho-behavioral comparisons between ballet dancers and their non-dancing peers is needed, particularly in Australia where published research of this nature is scant. Finally, while eating disorders (e.g., anorexia nervosa, bulimia nervosa) among dancers have been previously examined, the related condition of disordered eating has received far less attention by researchers (Wein & Micheli, 2002).
Therefore, the purposes of the present study were to compare ballet dancers and non-dancers on selected measures of disordered eating patterns, dysfunctional attitudes, and selected dispositions previously linked to eating disorders, and to determine the characteristics that predispose ballet dancers for developing disordered eating patterns. It was hypothesized that dancers would be at greater risk for developing eating disorders than non-dancers based on their higher scores for Drive for Thinness, Bulimia, Perfectionism, and Body Dissatisfaction subscales of the Eating Disorder Inventory-2 (EDI: Garner, 1990). It was also predicted that dancers would score higher than non-dancers on the three sub-scales of the Food Intake Attitude Survey (FIAS) generated for this study (discussed in the next section) which is focused on the condition of disordered eating.
Participants All female ballet students (N = 58) from four dance studios in Sydney, Australia, volunteered to participate in the study. Dance teachers verbally encouraged the dancers to complete the inventories used in the study when distributing the consent forms for their parent's signature. The criteria for inclusion into the non-dance control group was that participants were not currently competing in sport or regularly participating in classical, jazz, or tap dancing. The consent form explained the purpose of the study and allowed the option of their daughter not completing the inventories without any negative ramifications. Each participant signed a "Human Participant's Consent Form" indicating her willingness to volunteer for the study, and that she could withdraw from the study at any time without any negative consequences. To eliminate any possible teacher influence in the dancers responses, dance teachers were not present during data collection. All dancers were engaged in a full-time dancing training course at least 5 days per week. While socioeconomic status (SES) was not controlled in this study, this was a private program in an exclusive studio. Families of the participants would very likely be classified as moderate to high SES.
Only dancers whose parent signed the consent form participated in the study, a 100% return rate, although one dancer was absent on data collection day due to illness (N=57). The final sample consisted of female dancers ranging in age from 15.3 to 17.6 yrs (M = 17.8 yrs, SD = 2.11). The non-dance (control) group consisted of 51 female students from a private high school in Sydney located in the same school district as the dance group. Non-dancers ranged in age from 14.6 to 17.3 yrs (M = 17.1 yrs, SD = 1.12).
Eating Disorder Inventory (EDI). The EDI-2 (Garner, 1990) was used to identify characteristics and antecedents of eating disorders. Three of the subscales, Drive for Thinness, Bulimia, and Body Dissatisfaction assess attitudes and behaviors concerning eating, weight, and body shape. Five subscales, Ineffectiveness, Perfection, Interpersonal Distrust, Interoceptive Awareness, and Maturity Fears assess general psychological dispositions that are clinically relevant to eating disorders.
Consistent with Sundgot-Borgen (1994a, b), participants were determined to be at risk for eating disorders if their scores on the Drive for Thinness and Body Dissatisfaction subscales were elevated. For both subscales, an elevated score was defined as at or above the mean for a combined eating disordered sample (> 14.5 for DT, and > 16.6 for BD). This standard provided a conservative definition of at-risk participants and reduced false positives, because high scores on these subscales accurately characterize eating disordered patients (Gardner, 1990).
The criterion for each EDI-2 subscale was a coefficient of internal consistency above .80 for an eating disorder sample. Alpha coefficients for a combined eating disorder sample (N = 889) for the eight subscales ranged from .80 to .92. For a sample of non-patient college women (N = 268), alpha estimates for the Perfectionism (P) and the Maturity Fears (MF) scales were .79 and .80, respectively. Other subscale estimates included Interpersonal Distrust (ID) and Interoceptive Awareness (IA), r = .81, Bulimia scale, r = .82, Drive for Thinness (DT), r = .90, Ineffectiveness (I), r = .90, and Body Dissatisfaction (BD), r = .92. Similar internal consistency reliability estimates were found for a sample of 158 first and second year psychology students. The ability of the items to discriminate between eating disorder and non-patient samples is a source of evidence for criterion validity. A subsequent study has also found that the EDI subscales differentiate between eating disorder and non-clinical groups (Garner, Olmsted, Davis, Rockert, Goldbloom & Eagle, 1990). The EDI was also validated in a study by Schoemaker, van Strien, and Staak (1994) in a nonclinical population using transformed and untransformed responses. Thus, the psychometric properties of the EDI-2 have showed good reliability and validity.
Food Intake Attitude Survey (FAIS). The FIAS was developed for this study to assess attitudes and behaviors specifically relevant to disordered eating in dancers that are not included in the ED1-2. The FIAS included three subscales, Importance of Physique (IP), Determinants of Eating (DE) and Weight Control Behaviors (WB). The IP subscale examined a dancer's perception of the importance of a lean physique for dancing and for their advancement, the DE subscale investigated the concerns and influence of significant others on food intake, and the WB subscale identified behavioral indicants of eating disorders. Pretesting of this inventory among 22 first- year college students who had participated in at least 3 years of ballet lessons and considered themselves "highly skilled" indicated alpha's for each subscale of .71 (IP), .77 (DE), and .84 (WB). This indicated strong compatibility among items within each scale. Inter-scale correlations of .23 (IP-DE),. 19 (IP-WB), and .27 (DE-WB) revealed that each sub-scale was independent. Also included in this inventory was an open question that solicited the dancers' perceptions on the development of eating disorders. Face validity of the inventory was established based on a perusal of all items by two professional ballet instructors and two advanced ballet students who supported the relevance and content of each item.
Cronbach alphas were computed to determine the reliability of the items in measuring eating attitudes and behaviors. For the FIAS, a high alpha (r = 0.88) indicated good reliability for these items. Item total correlations were also acceptable. Item-total correlations were satisfactory. Items were grouped into three subscales, Importance of Physique (IP), Determinants of Eating (DE), and Weight Control Behavior (WB). Cronbach alphas (r) were computed to determine the reliability of each subscale. The IP subscale contained items such as "I think a lean physique is essential for a dancer's appearance," "I think a lean physique is essential for a quality dance performance," and "I think a lean physique is essential for selection to a dance company" (r = 0.73). The DE subscale contained items such as "I think food restriction determines body shape and weight," "I think exercise determines body shape and weight," 8 ("My dance teacher is strict or concerned about my weight," "My dance teacher influences my attitude and behavior about my food intake," "My parents/family are strict or concerned about my weight," "My parents/family influence my food intake," "My friends/peers are strict or concerned about my weight," "My friends/peers influence my attitude and behavior about my food intake," and "The media influences my feelings about my eating habits" (r = 0.82).
The WB subscale included "I am constantly dieting," "I go hungry in order to keep my weight down," "I currently use or have used laxatives, diuretics, or diet pills in order to keep my weight down or reduce my appetite," "I skip meals or fast in order to keep my weight down," "I avoid certain foods in order to keep my weight down," and "I now engage in or have engaged in self-induced vomiting in order to keep my weight down" (r = 0.89).
To obtain data from the non-dancers of comparable demographic characteristics, teachers of three senior (year 12) classes from a secondary school allowed questionnaires to be completed immediately prior to class time. This resulted in a 100% return rate. When visiting both studios, the researcher was introduced to the students before classes commenced and asked students to volunteer to complete the questionnaires during their break. Similar to the dance groups, classroom teachers were not present when data were obtained.
Differences between dancers and non-dancers on age, weight, and difference weight indices were statistically compared (see demographic data and t values in Table 1). The variable weight, which was measured by a research assistant, described the participant's present weight. The variable difference weight was established by subtracting the participant's present weight from their desired weight. Unpaired t-tests were conducted to compare dancers and non-dancers' age, weight, and weight. No significant difference between dancers and non-dancers was found for the variables age and difference weight, while significant differences were found for weight. In addition, dancers and non-dancers did not differ significantly in age. Although no difference was found in terms of difference weight, on average, both dancers (n = 52 of 57, 91.0%) and non-dancers (n = 29 of 51, 56.8%) expressed a desire to weigh less than their present weight. Only two dancers (0.04%) and 11 non-dancers (21.5%) were satisfied with their present weight, while none of the dancers and only 2 (0.04%) of the non-dancers wanted to be heavier.
Eating Disorder Inventory
To determine if dancers are more at risk for developing eating disorders than non-dancers, a chi-square analysis was performed. Based on the criterion used in previous studies (e.g., Garner, Olmsted, & Polivy, & Garfinkel, 1984), 10 (17.5%) of the dancers were identified as being at risk for developing an eating disorder, while the remaining 17 (33.3%) dancers were classified as not at risk. Among the non-dancers, three (.06%) were identified as being at risk, while 27 (53%) were identified as not at risk for developing an eating disorder.
A one-way MANOVA was conducted to compare dancers and non-dancers on a combination of EDI-2 subscales, Drive for Thinness (DT), Bulimia (B), Body Dissatisfaction (BD), Ineffectiveness (I), Perfectionism (P), Interpersonal Distrust (ID), Interoceptive Awareness (IA), Maturity/Fears (M-F). The MANOVA was significant, F (8, 99) = 6.75,p < 0.01, ES = .07. Subsequent univariate analyses (dfs = 1,106) revealed significant dance/non-dance differences on subscales DT (F = 7.07,p < 0.05), BD (F = 6.60,p < 0.05), and P (F = 4.06,p < 0.05). No significant dance/non-dance differences (ps > .05) were found on subscales B(F[1, 106) = 0.27, I (F[1, 106] = 0.38, ID(F[1, 106] = 0.66, IA (F = 0.75), and M-F (F = 0.16).
The variance in drive for thinness as a function of group membership was 11.4%. In addition, 10.7% of the variance for body dissatisfaction, and 6.9% of the variance in perfectionism was predicted by knowing group membership. A perusal of mean scores, illustrated in Table 2, indicate that dancers have greater drive for thinness and body dissatisfaction, and are more perfectionistic than non-dancers. Both dancers and non-dancers demonstrate similar levels of interpersonal distrust, interoceptive awareness, and maturity fears, and do not differ in terms of bulimic tendencies.
Because DT partially determines whether a participant is at risk for developing a disordered eating pattern, the DT subscale was correlated with the remainder of the EDI-2 subscales. For dancers, DT was highly and significantly correlated with P, r = 0.77,p < 0.05. The correlation between DT and P for non-dancers, r = 0.36, p > 0.05, was relatively weak and non-significant. The difference between dancers and non-dancers for DT and P correlations was statistically significant, z = 2.35, p < 0.05.
Cramer's Phi was used to examine the association between scoring above a combined eating disorder mean for both DT and P subscales. With dancers and non-dancers combined, Cramer's Phi indicates a good association, V = 0.54, which is statistically significant, [chi square](1, N = 57) = 16.61,p < 0.05. A participant who scores above the mean for DT is also likely to score above the P scale mean as well. For dancers, Cramer's Phi indicates a strong (V = 0.75), statistically significant, [chi square](1, n = 27) = 15.01, p < 0.05, association. This suggests that if a dancer is identified as being at risk for a disordered eating pattern by scoring above the mean on the DT subscale, then she is most likely to score above the mean on the P subscale. For non-dancers, Cramer's Phi indicates a weak V = 0.14, not statistically significant association, [chi square] (1, n = 30) = 0.54, p > 0.05. This indicated a weak association between DT and P subscales for non-dancers.
DT was moderately correlated with BD for both dancers, r = 0.49,p < 0.05, and non-dancers, r = 0.56,p < 0.05. With dancers and non-dancers combined, Cramer's Phi indicates a moderately strong association, V = 0.59. This suggests that a participant who is at risk for developing an eating disorder, will most probably score above the mean for body dissatisfaction as well. This is a statistically significant association, [chi square] (1, N= 57) = 19.61,p < 0.001. Separate analyses indicate a strong association for dancers, V = 0.64, which is statistically significant, [chi square](l, n =27) - 10.92,p < 0.001. There is a moderate association for non-dancers, V = 0.43, which is also significant, [chi square] (1, n = 30) = 5.l, p < 0.01. These results suggest that being at risk for an eating disorder was significantly associated with body dissatisfaction for both dancers and non-dancers.
In addition, the DT subscale was moderately correlated with IA for both dancers, r = 0.49,p < 0.05, and non-dancers, r = 0.41,p < 0.05. Therefore, pursuit of thinness was moderately related to the ability to recognize internal states for both dancers and non-dancers. The correlation between DT and IA was stronger for non-dancers, r = 0.51, p < 0.05, than for dancers, r = 0.39,p < 0.05. The pursuit of thinness was moderately associated with feelings of ineffectiveness in non-dancers, and only moderately associated with ineffectiveness in dancers.
Because it was of interest to compare dancers and non-dancers on each of the three subscales, unpaired t-tests, rather than MANOVA, were computed. A Bonferroni technique was applied to control for Type I error, with alpha set at p < .01. Dancers (M = 3.86, SD = 1.31) markedly differed from non-dancers (M = 2.06, SD = 1.57) on the IP scale, t (106) = 3.89, p < .01). Perhaps not surprisingly, dancers perceive their level of fitness and body shape significantly more important than their non-dancer peers. Comparisons between groups on the DE measure indicated that dancers (M = 3.42, SD = 1.71) scored markedly higher than non-dancers (M = 2.02, SD = 1.66), t (106) = 3.12, p < .02. Apparently dancers' eating habits are more influenced by significant others (e.g., dance teacher, friends) and the links between eating and exercise to body shape and weight than non-dancers. Finally, dancers (M = 4.17, SD = 1.76) differed significantly from non-dancers (M = 2.27, SD = 1.04) on the WB scale, t (106) = 4.32,p < 0.001. This indicated that dancers were engaged in weight control behaviors, and with greater intensity to a greater extent than the non-dancers in this study.
The variety of weight control behaviors and the frequency with which they are used by dancers and non-dancers is illustrated in Table 3. The frequency of each method was derived by summing the percentage of dancers and non-dancers who rated that they often, usually, or always used the behavior for weight reduction. This was thought to represent a subgroup of dancers and non-dancers who regularly used each behavior. Some of the behaviors are innocuous, such as avoiding certain foods. Others have been characterized as pathogenic, including the use of laxatives, diuretics, diet pills, and self-induced vomiting (Rosen, McKeag, Hough, & Curley, 1986).
Pearson product moment correlation was performed to examine relationships between WB and EDI subscales. WB was highly correlated with DT for dancers, r = 0.77, p < 0.01, and moderately correlated with DT for non-dancers, r = 0.52, p < 0.05. These results support the hypothesis of a relationship between the pursuit of thinness and engaging in weight control behaviors. In addition, for dancers, WB was moderately correlated with the P subscale, r = 0.64, p < 0.02. The relationship between WB and P supports the hypothesis that perfectionism would be associated with the use of weight control behaviors in dancers. For non-dancers, WB correlated weakly and non-significantly with all other subscales of the EDI-2.
Pearson product-moment correlations were computed on FIAS and EDI-2 subscales. WB correlated highly with DE for dancers, r = 0.76, p < 0.001, and non-dancers, r = 0.91, p < 0.001. This suggests that a relationship exists between weight control behaviors and the concern and influence of significant others. WB was also moderately correlated with IP, r = 0.57, p < 0.05, suggesting that weight control behavior is related to the attitude of the dancer as to the importance of a lean physique in dance.
The purposes of the present Australian study were to examine the eating attitudes, behaviors, and selected dispositional profiles of ballet dancers linked to disordered eating patterns, as opposed to non-dancers, and to determine the factors that most likely predisposed ballet dancers at risk for developing a preoccupation with body shape and perfectionism. Three characteristics were identified, the type and frequency of weight control behaviors used by dancers, the association between the dancers' concerns about their diet and weight, and evidence of psychological disturbance among the dancers that are commonly found in eating disordered patients.
The results of this study supported the hypothesis that dancers would have a greater pursuit of thinness and body dissatisfaction than non-dancers. Further, as predicted, dancers were more perfectionistic than non-dancers. By contrast to expectations, however, dancers did not demonstrate more bulimic tendencies than non-dancers. Finally, the absence of significant differences on the remaining EDI scales supported the hypothesis that dancers would not have similar characteristics as individuals with psychopathological eating disorders. In particular, similar scores between dancers and non-dancers on subscales I, ID, IA, and Maturity Fears subscales of the EDI supports the contention that concerns about having a lean physique and maintaining a proper diet and weight are normative among dancers. Therefore, adolescent-aged dancers and non-dancers may experience similar feelings without the psychological disturbance commonly associated with eating disordered patients (Brooks-Gunn et al., 1988).
The dancers in the present study weighed significantly less than the non-dancers, and only 7% of the dancers in this study were satisfied with their present weight, while the other dancers preferred to weigh less. These findings are consistent with previous studies indicating that dancers are lighter than non-dancers (Clough & Wilson, 1993; Davis, 1992).
Another finding was that dancers were more at risk for developing eating disorders than non-dancers. This finding may be partly explained by the propensity of dancers to exercise addiction. Petrie (1993) and Martin and Hausenblas (1998) found eating disorder symptomatology in female collegiate gymnasts and among female aerobic instructors, respectively. Both groups indicated a psychological commitment to exercise. The drive for thinness through excessive exercise has been termed activity anorexia (Dess, 2000). The prevalence of eating disorders is especially higher among athletes in "aesthetic" and "weight-dependent" sports, including dancers (Sundgot-Borgen, 1994b).
As predicted in the present study, dancers scored higher on the DT and BD subscales than non-dancers. It is feasible to surmise that dancers are excessively concerned about dieting and weight and dissatisfied with their bodies, which is consistent with previous studies (e.g., Druss & Silverman, 1979; LeGrange, Tibbs, & Noakes, 1994). It is also possible that eating disorders are over-represented in ballet dancers because dancers need to focus increased attention on, and control over, their body shape in order maintain a lean physique (Garner & Garfinkel, 1980; LeGrange et al., 1994). In addition, ballet, as a form of physical activity, is an inefficient method of burning calories. Therefore, attempts to control weight and maintain a lean physique are achieved far more effectively through dietary restraint, a practice that has been implicated in the development of eating disorders (Calabrese et al., 1983; Cohen, Segal, Witriol, & McArdle, 1982; Hill, 1994).
The finding of greater body dissatisfaction among dancers in this study might be partially explained by sociocultural factors. Although dancers and non-dancers, especially females, experience similar societal pressures to be slim, dancers are further pressured by expectations of a lean physique (Garner & Garfinkel, 1980). In addition, not unlike female athletes who define their identity by the quality of their athletic performance (Thompson, 1987), many dancers and dance teachers also believe that thinness is associated with better dance performance. The pursuit of thinness, therefore, may be a vehicle to improve a dancer's self-concept. This thesis awaits further research.
The finding that dancers scored higher than non-dancers on perfectionism is consistent with previous research that indicates that dancers are motivated by a drive to achieve perfection (Clough & Wilson, 1993; Druss & Silverman, 1979). To Clough and Wilson, participation in dance encourages perfectionistic attitudes. Furthermore, normal, as opposed to neurotic, perfectionism often enables individuals, dancers and non-dancers, to succeed (Hewitt & Flett, 1991; Sundgot-Borgen, 1994a). Therefore, it is likely that the high self-expectations and perfectionistic attitudes of dancers may facilitate certain behavioral tendencies that enhance the likelihood of success in a highly competitive dance world.
The significant relationship between DT and P among dancers found in this study supported previous findings of a strong relationship between reported eating disorder characteristics and perfectionism by Clough and Wilson (1993). According to the researchers, prominent eating problems tend to occur in individuals who have unrealistic high expectations and self-critical expectations. Garner and Garfinkel (1980) contend that the combination of high self-expectations among dancers, coupled with an environment that extols the virtues of thinness, places dancers at risk for developing eating disorders. Taken together, these factors could partially explain the relationship between DT and P for dancers but not for non-dancers in the present study.
In support of the present results, Harris and Greco (1990) and Warren, Stanton, and Blessing (1990) found that adolescent female gymnasts and female non-athletes scored similarly on the Bulimia subscale. However, because bulimia nervosa is more prevalent in athletes and dancers than non-athletes and non-dancers (e.g., Davis, 1992; Petrie & Stover, 1993), it would be interesting to determine, in future research, if eating disorders are underreported using the EDI. Still, the dancers in the present study did not appear to demonstrate eating disorder psychopathology. As indicated earlier, although dancers and other individuals with eating disorders share similar weight and dieting concerns, dancers tend to lack the associated psychopathology of eating disorders (Garner et al., 1984; Holderness et al., 1994). The profiles of the dancers in the present study reflect what Garner et al. (1984) have described as benign, that is, women who score high on the DT, BD, and P subscales of the EDI, and low on all other subscales. This profile indicates that the pursuit of thinness is not always associated with psychopathology.
The hypothesis that dancers would engage in more weight control behaviors compared to non-dancers was supported in this study. These results are consistent with previous studies (e.g., Clough & Wilson, 1993; LeGrange et al., 1994). According to Druss and Silverman (1979), although most dancers possess a lean physique, they continue to diet or to employ other means to become even thinner.
One inherent limitation to a study of this nature is the reliance on self-report measures. It is well known that underreporting is an inherent limitation of eating disorders research. Studies that rely on self-disclosure of potentially embarrassing or inappropriate behavior are increasingly vulnerable to distortions due to response bias and inaccurate reporting (Rosen & Hough, 1988; Stoutjesdyk & Jevne, 1993). Sundgot-Borgen (1994b) has suggested that due to the secretive nature of eating disorders, dancers may be reluctant to respond truthfully to the questionnaires. Reliance on volunteers from the respective groups in this study represents another source of potential response bias. It is possible that individuals who choose not to participate have different eating attitudes and behaviors than volunteers do.
In summary, the primary findings were that the EDI indicated group differences on the perfectionism measure and that dancers were less satisfied with their body, and, therefore, pursue more eating strategies that promote more thinness, as compared to non-dancers. Examining the extent to which these variables contribute to eating disorders awaits further research. In addition, a more extensive investigation is necessary to determine whether dancers demonstrate psychopathology common in patients with eating disorders, and if not, whether they remain at risk for developing eating disorders compared to female adolescents of a similar age. Relationships between eating disorders and selected dispositions such as neurotic perfectionism, fear of failure, low self-concept, and a dysfunctional (negative) addiction to exercise also await further research.
Table 1. Means, standard deviations, and t values (df=106) for age, weight and difference weight for dancers and non-dancers. Dancers (n = 27) Non-dancers (n = 30) Mean SD Mean SD t Variable Age 17.30 2.46 17.53 0.68 -0.48 Weight 51.22 6.65 56.17 5.61 -3.05 * Diff Weight -3.72 1.89 -2.67 3.51 -1.43 * p<.01 Table 2. Means and standard deviations for the eight scales of the EDI for dancers and non-dancers Dancers (n = 57) Non-dancers (n = 51) Mean SD Mean SD Variable Drive for Thinness 9.26 3.75 4.93 2.53 Bulimia 2.56 1.51 2.03 1.03 Body Dissatisfaction 16.56 2.57 11.53 4.44 Interoceptive 4.00 3.13 3.40 2.07 Perfectionism 7.00 2.51 4.63 2.35 Interpersonal Distrust 2.70 2.37 3.43 3.42 Interoceptive Awareness 4.26 1.91 5.47 3.53 Maturity/Fears 5.33 1.31 4.87 3.45 Table 3. Weight control methods used by dancers and non-dancers to reduce or maintain weight. % of dancers % of non-dancers using using Method (n = 57) (n = 51) Constant dieting 29.6 13.3 Laxatives/diuretics/diet pills 11.1 6.7 Skipping meals/fasting 25.9 10.0 Avoiding certain foods 62.9 1.5 Self-induced vomiting 14.8 6.6
The author expresses his appreciation to Ms. Allison Scott for her contribution in the collection of selected data for this study. A copy of the inventories used in this study is available from the author (e-mail: email@example.com)
American Psychiatric Association (1987). Diagnostic and Statistical Manual of Mental Disorders. Third Edition revised (DSM-III-R). Washington, DC: American Psychiatric Association.
Brooks-Gunn, J., Burrow, C., & Warren, M. (1988). Attitudes toward eating and body weight in different groups of female adolescent athletes. International Journal of Eating Disorders, 7, 749-757.
Brooks-Gunn, J., Warren, M., & Hamilton, L. (1987). The relation of eating problems and amenorrhea in ballet dancers. Medicine and Science in Sport and Exercise, 19, 41-44.
Brownell, K., Rodin, J., & Wilmore, J. (1992). Eating, body weight and performance in athletes: Disorders of modern society. Philadelphia, London: Lea & Febiger.
Bruch, H. (1973). Eating disorders: Obesity, anorexia nervosa, and the person within. New York: Basic Books, Inc.
Bruch, H. (1978). The golden cage. Cambridge, MA: Harvard University Press.
Calabrese, L., Kirkendall, D., Floyd, M., Rapoport, S., Williams, G., Weiker, G., & Bergfeld, J. (1983). Menstrual abnormalities, nutritional patterns and body composition in female classical ballet dancers. The Physician and Sportsmedicine, 11, 86-98.
Clough, M., & Wilson, B. (1993). The relationship between eating disorder characteristics and perfectionism among 22 dance students 17-21 years. Sport Health, 11, 5-8.
Cohen, J., Potosnak, L., Frank, O., & Baker, H. (1985). A nutritional and hematologic assessment of elite ballet dancers. The Physician and Sportsmedicine, 13, 43-52.
Cohen, J., Segal, K., Witriol, I., & McArdle, W. (1982). Cardiorespiratory responses to ballet exercise and the VO2 max of elite ballet dancers. Medicine and Science in Sports and Exercise, 14, 212-217.
Comer, R.J. (2001). Abnormal psychology (4th ed.). New York: W.H. Freeman.
Davis, C. (1992). Body image, dieting behaviors, and personality: A study of high-performance female athletes. International Journal of Sport Psychology, 23, 179-192.
Dess, N.K. (May/June 2000). Killer workout: The dark side of diet and exercise. Psychology Today, 32, 26.
Druss, R., & Silverman, J. (1979). Body image and perfectionism of ballerinas: Comparison and contrast with anorexia nervosa. General Hospital Psychiatry, 1, 115-121.
Fay, M., & Price, C. (1993). The shape and weight of advertising models, and the prevalence of eating disorders among young women. Proceedings of the Nutrition Society of New Zealand, 18. 59-74.
Garner, D.(1990). The Eating Disorder Inventory professional manual-2. Odessa: Psychological Assessment Resources.
Garner, D., & Garfinkel, P. (1980). Socio-cultural factors in the development of anorexia nervosa. Psychological Medicine, 10, 647-656.
Garner, D., Olmsted, M., Davis, R., Rockert, W., Goldbloom, D., & Eagle, M. (1990). The association between bulimic symptoms and reported psychopathology. International Journal of Eating Disorders, 9, 1-15.
Garner, D., Olmsted, M., Polivy, J., & Garfinkel, P. (1984). Comparison between weight-preoccupied women and anorexia nervosa. Psychosomatic Medicine, 46, 255-265.
Hamilton, L., Brooks-Gunn, J., & Warren, M. (1985). Sociocultural influences on eating disorders in professional female ballet dancers. International Journal of Eating Disorders, 4, 465-477.
Hamilton, L., Brooks-Gunn, J., & Warren, M. (1986). Nutritional intake of female dancers: A reflection of eating problems. International Journal of Eating Disorders, 5, 925934.
Harris, M., & Greco, D. (1990). Weight control and weight concern in competitive female gymnasts. Journal of Sport and Exercise Psychology, 12, 427-433.
Hausenblaus, H.A., & Carron, A.V. (1999). Eating disorder indices and athletes: An integration. Journal of Sport and Exercise Psychology, 21,230-258.
Hergenroeder, A., Fiorotto, M., & Klish, W. (1991). Body composition in ballet dancers measured by total body electrical conductivity. Medicine and Science in Sports" and Exercise, 23, 528-533.
Hewitt, P., & Flett, G. (1991). Perfectionism in the self and social contexts: Conceptualization, assessment, and association with psychopathology. Journal of Personality and Social Psychology, 60, 456-470.
Hill, A. (1994). The nature and consequences of pre-adolescent dieting. British Nutritional Foundation, Nutritional Bulletin, 19, 95-103.
Holderness, C., Brooks-Gunn, J., & Warren, M. (1994). Eating disorders and substance abuse:A dancing versus a nondancing population. Medicine and Science in Sport and Exercise, 26. 297-302.
Johnson, A., Steinberg, R., & Lewis, W. (1988). Bulimia. In K. Clark., R. Parr, & W. Castelli (Eds.), Evaluation and management of eating disorders, anorexia, bulimia and obesity. Champaign, Illinois: Life Enhancement Publications.
Lancelot, C., Brooks-Gurm, J., Warren, M., & Newman, D. (1991). Comparison of DSM-III and DSM-III-R bulimia nervosa classifications for psychopathology and other eating behaviors. International Journal of Eating Disorders, 10, 57-66.
LeGrange, D., Tibbs, J., & Noakes, T. (1994). Implications of a diagnosis of anorexia nervosa in a ballet school. International Journal of Eating Disorders. 15, 369-376.
Maloney, D. (1983). Anorexia nervosa and bulimia in dancers--acute diagnosis, treatment and planning. Clinics in Sports & Medicine, 2, 549-555.
Martin, K.A., & Hausenblas, H.A. (1998). Psychological commitment to exercise and eating disorder symptomatology among female aerobic instructors. The Sport Psychologist, 12, 180-190.
O'Dea, J. (2002). Can body image education programs be harmful to adolescent females? Eating Disorders: Journal of Treatment & Prevention, 10, 1-13.
Petrie, T. (1993). Disordered eating in female collegiate gymnasts: Prevalence and personality/attitudinal correlates. Journal of Sport and Exercise Pyschology, 15, 424-436.
Petrie, T., & Stoever, S. (1993). The incidence of bulimia nervosa and pathogenic weight control behaviors in female collegiate gymnasts. Research Quarterly for Exercise and Sport, 64. 238-241.
Pierce, E., Daleng, M., & McGowan, R. (1993). Scores on exercise dependence among dancers. Perceptual and Motor Skills, 76, 531-535.
Rosen, L., & Hough, D. (1988). Pathogenic weight control behaviors of female college gymnasts. The Physician and Sportsmedicine, 16, 141-145.
Rosen, L., McKeag, D., Hough, D., & Curley, V. (1986). Pathogenic weight control behavior in female athletes. The Physician and Sportsmedicine, 14, 79-86.
Schoemaker, C., van Strien, T., & Staak, C. (1994). Validation of the Eating Disorder Inventory in a nonclinical population using transformed and untransformed responses. International Journal of Eating Disorders, 15, 387-393.
Slavin, J. (1987). Eating disorders in athletes. Journal of Physical Education Recreation and Dance, 58, 33-36.
Sundgot-Borgen, J. (1994a). Eating disorders in female athletes. Sports Medicine, 17, 176-188.
Sundgot-Borgen, J. (1994b). Risk and trigger factors for the development of eating disorders in female athletes. Medicine and Science in Sport and Exercise, 26, 414-419.
Thompson, R. (1987). Management of the athlete with an eating disorder: Implications for the sport management team. The Sport Psychologist. 1, 114-126.
Warren, B., Stanton, A., & Blessing, D. (1990). Disordered eating patterns in competitive female athletes. International Journal of Eating Disorders, 9, 565-569.
Wein, D., & Micheli, L. (2002). Nutrition, eating disorders, and the female athlete triad. In D.L.
Mostofsky & L.D. Zaichkowsky (Eds.), Medical and psychological aspects of sport and exercise (pp. 91 - 111). Morgantown, WV: Fitness Information Technology.
Woodside, D., & Garfinkel, P. (1989). An overview of the eating disorders: Anorexia Nervosa and Bulimia Nervosa. National Institute of Nutrition Review. 8, 1-4.
Mark H. Anshel
Middle Tennessee State University
Address Correspondence To: Mark Anshel, Ph.D.,Professor, Department of Health, Physical Education, Recreation, & Safety, Middle Tennessee State University, Murfreesboro, TN 37132…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Sources of Disordered Eating Patterns between Ballet Dancers and Non-Dancers. Contributors: Anshel, Mark H. - Author. Journal title: Journal of Sport Behavior. Volume: 27. Issue: 2 Publication date: June 2004. Page number: 115+. © 1999 University of South Alabama. COPYRIGHT 2004 Gale Group.
This material is protected by copyright and, with the exception of fair use, may not be further copied, distributed or transmitted in any form or by any means.