Karlson, Kristine A. MD*; Becker, Carolyn Black PhD†; Merkur, Amanda‡
While the prevalence of eating disorders varies widely depending on the population examined, high rates have been documented in female athletes, 1–3 particularly those involved in weight-related sports. 2,4,5 Weight-related sports with documented eating disorder issues include distance running, 2 which rewards low weight with improved performance, and aesthetic sports such as figure skating, 6 gymnastics, 7 and ballet, 8 which value appearance. Weight-related sports also include weight-dependent sports, which require athletes to “make weight” for competition, such as wrestling, judo, and lightweight rowing.
Disordered eating has not been well investigated in female athletes participating in a relatively new weight-dependent sport for women—lightweight rowing. This may be due in part to the fact that lightweight rowing for women, which uses a single weight class of 130 pounds at the collegiate level, has traditionally involved small numbers of participants. Recently, Title IX, which requires equal participation for women in collegiate sports in the United States, has yielded a vast increase in heavyweight (no weight class) rowing opportunities for women at the collegiate varsity level, as well as interest in lightweight women's rowing. Still, lightweight rowing for women has grown slowly, partly due to the perception among coaches and administrators that adding lightweight rowing programs for women and having women “make weight” could increase the prevalence of eating disorders. Such fears have largely been based on comparisons made between lightweight rowing and wrestling, a weight-dependent sport shown to be associated with disordered eating in men. 9
There are a number of parallels between wrestling and lightweight rowing. Both sports require participants to reach a predetermined weight. Wrestlers frequently lose large amounts of body weight during the season and often use unhealthy methods to do so. 9 In 1993, Thiel 10 found an equal prevalence of disturbed body image in lightweight male rowers and wrestlers. Preliminary data suggests that, similar to wrestlers, lightweight women rowers are also at risk for using unhealthy methods to make weight, such as drastic dieting, purgatives, and dehydration (K. A. Karlson, unpublished observations). Wrestling frequently requires that participants attempt to reach a body weight that is unnaturally low for their build. As lightweight rowing becomes more competitive, particularly if it involves competition for scholarships, larger women may try to make weight, increasing the likelihood of disordered eating behaviors when they attempt to reach a weight far below their natural weight. Results from one study 5 found that 37% of athletes with anorexia nervosa reported that the main reason for the development of their eating disorder was dieting and weight fluctuations required for their sport, suggesting that making weight may predispose athletes to the development of eating disorders.
There may be significant differences, however, between lightweight rowing, wrestling, and other weight-dependent sports. A study of wrestlers found that eating concerns were not carried into the off season, 11 suggesting that their use of disordered eating practices did not contribute to the development of eating disorders. This may also be true for rowers, who make weight in similar ways. Unlike wrestling, lightweight collegiate women rowers have only one weight class, 130 pounds. There are no pressures to lose more than required to make that weight class, also unlike sports that emphasize body appearance. Such additional weight loss is actively discouraged as rowing is also a strength-dependent sport, and athletes weakened by dehydration or excessive weight loss will not perform well. The presence of a single weight class may provide a protective factor if women who enter lightweight rowing preselect on the basis of being able to maintain the required weight with relative ease.
There are few previous studies examining lightweight women rowers. Three studies have compared lightweight rowers to heavyweights, who have no weight restrictions. Terry and Waite 12 examined eating attitudes in an elite British group among both men and women and found that 19.4% of lightweight women, but no heavyweight women, scored above the critical level on the Eating Attitudes Test (EAT) with 31 athletes in each group. A second study 13 found that 21% of elite female lightweight rowers scored over the threshold for eating disorder concerns, compared with 6% of male lightweight rowers and 4% of female heavyweight rowers. A third study, 4 however, found only a nonsignificant trend toward differences between lightweight and heavyweight rowers in weight practices and disordered eating in an American collegiate population. This discrepancy may be due to the appropriate self selection of leaner and smaller women to row lightweight at the collegiate level, while at the elite level there are increased pressures for larger women to attempt to make weight. One unrelated small study 14 looked for permanent metabolic changes related to repeatedly losing and regaining weight, and did not find long-term detrimental health consequences as measured by thyroid function. A final study 15 looked at variations in body composition between the preseason and competition periods and found a significant reduction in body fat but no change in fat-free mass. These elite athletes did not report use of pathologic weight-making practices. A limitation of these previous studies has been their relatively small sample size. In addition, only one study 4 has examined eating behaviors in collegiate versus elite women rowers, and the sample size was limited (lightweight women, n = 17).
The purpose of the current study was to examine the prevalence of eating disorders and related behaviors in a larger sample of collegiate lightweight women rowers as compared with distance runners, who historically have an increased prevalence of eating disorders, and collegiate controls. It was hypothesized that since there is a set weight for the rowers and disincentive to lose additional weight, lightweight rowers would display an increase in weight-loss behaviors, but they would not significantly differ from controls in the presence of probable eating disorders. In contrast, it was hypothesized that distance runners, with ongoing pressures for thinness, would not only show an increased rate of disordered behaviors but would also exhibit an increased rate of probable eating disorders.
Participants were 122 collegiate lightweight women rowers (excluding coxswains), 79 collegiate women runners competing at distances of 800 meters or greater, and 95 collegiate women recruited as control subjects. The lightweight rowers were recruited at two East Coast end-of-season championship regattas, and represented 91% of the lightweight women competitors at those regattas. All rowers who were approached consented to participate. Runners were recruited from a collegiate track team and at the 1998 New England regional collegiate track championships. Due to the nature of track meets, it was not possible to approach all runners at convenient times (i.e., it was not possible to determine whether a runner had completed competing). All runners approached agreed to participate. A number of runners, however, requested to complete the questionnaire at a later time after they had finished competing. Unfortunately all such runners were not successfully relocated as some of the teams had departed before being relocated. As such, it is not possible to determine exactly what percentage of runners at the meet completed the study. Numbers attending were not available. The control subjects were recruited from a psychology class and at a college dining hall, where there was one refusal. The control group included varsity athletes from other sports, some of whom were heavyweight rowers. Participation was voluntary and confidential, with participants giving verbal consent to participate.
The Eating Disorders Examination-Questionnaire (EDE-Q)
The EDE-Q 16 is a self-report version of the Eating Disorders Examination (EDE), a semistructured interview that is considered to be the “gold standard” in the assessment of eating disorders. 17 Like the EDE, the EDE-Q assesses the main behavioral and attitudinal features of eating disorders over a 28-day period. The EDE-Q has been directly compared with the EDE in three published reports, 16,18,19 and has consistently performed well. In addition, Beglin and Fairburn 20 found that a screening version of the EDE-Q was better at case identification than the EAT-26, another commonly used eating disorder screening questionnaire. More recently Luce and Crowther 21 examined the internal consistency and test–retest reliability of the EDE-Q. In their analysis, Cronbach alphas were used at two different time periods to assess internal consistency. Alphas ranged from 0.78 and 0.81 for the eating concern subscale to 0.92 and 0.93 for the shape concern. Pearson correlations between time 1 and time 2 for the subscales ranged from 0.81 to 0.94 and were all highly significant. These results further support the psychometric adequacy of this measure.
In addition to generating a total score and screening for eating disorder cases, the EDE-Q produces four subscale scores, Eating Concern, Shape Concern, Weight Concern, and Restraint. Subscale items are rated on a 7-point scale either in terms of frequency (e.g., “On how many days out of the past 28 days have you been consciously trying to restrict the amount of food you eat to influence your shape or weight?”), or severity (e.g., “Over the past 28 days has your weight influenced how you think (judge) yourself as a person?”). Subscales are calculated by summing the scores across the appropriate items and then dividing by the total number of items included in the subscale. Subscale scores range from 0 to 6, with higher numbers indicating greater degrees of disturbance.
In addition to completing the EDE-Q, participants also provided their height, current weight, highest and lowest weight in the past 3 months and past year, and a detailed menstrual history including age at menarche, number of periods in the past year, and use of oral or injectable contraceptives. Menstrual history data were used to verify the accuracy of reported secondary amenorrhea, as contraceptive use can mask otherwise clinically important amenorrhea.
Results are reported in terms of means and standard deviation and were analyzed using analyzes of covariance (ANCOVA). Post hoc differences were determined using simple contrast analyses.
Demographics are presented in Table 1, and mean subscale scores are presented in Table 2. Since the control group contained a number of varsity athletes from non-weight-concerned sports, the control group was divided into two groups on the basis of varsity sports participation. Results from t test analyses indicated that there were no significant differences between these two groups on subscale scores. Thus, the original combined control group was used in all further analyses. There were significant differences between the three groups on both age [F(2,293) = 7.18, p < 0.0001] and body mass index (BMI) [F(2,271) = 30.24, p < 0.0001]. Age and BMI were therefore controlled for in all subsequent analyses.
Results from separate one-way ANCOVA, with age and BMI used as covariates, revealed significant differences between groups on both the Restraint [F(2,268) = 3.10, p < 0.05] and Shape Concern [F(2,266) = 5.47, p < 0.01] subscales. Contrast analyses indicated that rowers were significantly more restrained than both the runners and control group (p < 0.05). In contrast, on the Shape Concern subscale, rowers were significantly less concerned than either runners or control subjects. Runners and control subjects did not significantly differ from each other on either subscale. There were no significant differences between the three groups on the remaining subscales (i.e., Eating Concern, Weight Concern) or on the Total Score.
To further explore the possible impact of the weight requirement in rowing, three specific problematic weight-control strategies (vomiting, laxative abuse, and diuretic abuse) were examined across the three groups (see Table 3). Chi-squared (χ2) analyses showed a significant difference between groups for diuretic misuse, with more rowers reporting diuretic abuse than either runners or controls. There were no significant differences between groups on laxative misuse and vomiting, although rowers endorsed laxative misuse and runners endorsed vomiting to a greater degree than the other two respective groups.
Analyses of weight fluctuation indicated a significant difference between groups [F(2,252) = 12.77, p < 0.0001], with runners (M = 6.6 lbs., SD = 5.0) reporting significantly less weight fluctuation over a 1-year period as compared with rowers (M = 10.6 lbs., SD = 5.5) and control participants (M = 8.8 lbs., SD = 4.7). Rowers, however, showed less lifetime weight fluctuation than runners. As seen in Table 1, the mean weight for the rowers was 127.97 lbs. (SD = 3.39), and their highest mean reported weight since age 16 was 139.1 lbs (SD = 9.67).
Finally, the prevalence of probable eating disorders in each group was examined (see Table 4). To identify probable eating disorder cases, responses on the main diagnostic questions along with responses from additional nonoverlapping key questions identified by Beglin and Fairburn, 20 such as guilt about shape and weight and interference with concentration due to thoughts about food, were analyzed. In the complete sample, 26 probable eating disorder cases were identified, according to criteria from the Diagnostic and Statistical Manual of Mental Disorders (DSM IV). 22 Three participants met full criteria for bulimia nervosa, and no participants met full criteria for anorexia nervosa. The remaining 23 cases met DSM IV criteria for probable Eating Disorder–Not Otherwise Specified (ED-NOS). Of these 23 cases, 3 met criteria for subthreshold anorexia nervosa secondary to not meeting the weight or menstruation criteria. Of the remaining 20 ED-NOS cases, 15 endorsed repeated objective bulimic episodes within the past month, along with high weight dissatisfaction and ratings of 4 or higher on importance of weight or shape. Several of these participants also endorsed using vomiting, laxatives, or diuretics to control their weight. All nonpurging ED-NOS cases reported high levels of vigorous exercise designed to influence weight or shape. The additional 5 ED-NOS cases included 3 participants who met all of the weight concern criteria listed above and who reported very frequent subjective bulimic episodes (12 or more in the past month), and 2 very weight-dissatisfied participants who regularly used vomiting to manage their weight. There were no significant differences between groups in any of the eating disorder diagnostic categories.
The present study represents the largest examination to date of potential eating disordered behavior in lightweight women rowers. Despite concerns that participation in lightweight rowing might increase the prevalence of eating disorders, rowers in this collegiate sample did not show a higher rate of probable eating disorder cases. These results are consistent with those of Sykora et al., 4 who found only a nonsignificant trend in increased eating pathology between heavy and lightweight collegiate women rowers. These results, however, may not generalize to elite rowers, as the findings reported by Terry and Waite 12 suggest that elite lightweight women rowers may be at increased risk for developing eating disorders.
This lack of identification of probable cases of anorexia nervosa and bulimia in this collegiate athlete population is consistent with previous investigations. Johnson et al. 23 reported a prevalence of 1.1% meeting DSM IV criteria for bulimia and none meeting criteria for anorexia in a sample of 562 collegiate female athletes. However, in that sample, like the current study, there were a concerning number of female athletes displaying subthreshold disordered eating behavior. At the elite level, Sundgot-Borgen 5 found 1.3% of 522 female athletes met criteria for anorexia nervosa and 8% met criteria for bulimia. Similarly, Smolak 24 found in a metaanalysis that elite athletes appear to be at a higher risk than younger and lower level athletes and controls.
These results are also consistent with previous studies examining the prevalence of eating disorders in the general population. While it is difficult to establish norms due to variations in subpopulations and the inconsistencies of self-reported data, it has been reported that the prevalence of anorexia nervosa in the general population is 0.28% and the prevalence of bulimia is 1%. 25 In the collegiate and adolescent population, the prevalence is higher, as reported by Krahn, 26 in which 1.6% of 1,796 teen girls met DSM IIIR criteria for bulimia. In their review, Fairburn and Beglin 27 found the mean for bulimia prevalence varied between 2.6% and 9%, depending on how stringently the diagnostic criteria were used in individual studies. Establishing the prevalence of eating disorders, not otherwise specified, is more difficult, though Steiner and Lock 28 suggest it is probably significantly more common in young women than in the general population. One prospective study 29 found a 4% rate of development of an eating disorder in a 4-year study of adolescent girls. Finally, Johnson 23 found a rate of 9.2% reporting subthreshold disordered eating behavior in their study of collegiate athletes. The current study compares favorably with these data, suggesting that this age group is a population at risk, but that the prevalence of disordered eating in the rowers studied was not substantially higher than others in the same age group.
Although there were no significant differences in the overall prevalence of probable eating disorder cases, rowers, as hypothesized, showed significantly more eating restraint. This was not unexpected given that the rowers were surveyed on race day and were asked about their behaviors during the past 28 days, a period of time during which they would be expected to be trying to lose weight. One could speculate that this may have increased the level of restraint reported, and that rowers might not differ from other groups during noncompetitive times of year. While rowers reported increased levels of restraint, it is important to note that rowers' level of restraint does not appear to reach clinical significance. Although published norms for the EDE-Q are not, to our knowledge, available at this time, subscale scores for the EDE-Q have been found to be comparable to those of the EDE in direct comparisons. 16,18 Reported norms for the EDE restraint subscale range from 3.14 (SD = 1.22) to 3.7 (SD = 1.7) for bulimic and anorexic samples, 30 and 3.15 (SD = 0.33) for restrained controls. The mean of 2.01 found in this sample of rowers is higher than that found in normal controls (M = 0.79–0.91), but lower than those seen in clinical samples.
Although the rowers showed higher levels of restraint, they showed significantly less concern about shape than either runners or controls. One possible explanation for this finding is that the control participants were unusually concerned about their shape. EDE norms support this interpretation in that both runners and controls reported more shape concern than normal controls in previous studies. Runners and controls, however, also reported less shape concern than seen in clinical samples. 30 Rowers shape-concern subscale mean was between that of normal controls and restrained eaters. An alternative explanation of the discrepancy between the restraint and shape-concern subscale scores in rowers is that rowers were actively restricting their eating due to the demands of their sport, rather than due to concerns about body image. This interpretation appears consistent with the lower prevalence of overall probable eating disorder cases.
Rowers reported a mean maximum weight since age 16 of 139 pounds, while runners and controls reported mean maximum weights of 146 and 141 pounds, respectively. This suggests that these rowers may be naturally fairly close to the required racing weight of 130 pounds, which might decrease the likelihood of eating disorders. Rowers also reported larger weight fluctuations within the past year than runners. Larger weight fluctuations throughout the year might be expected for the rowers, who may actively lose weight during the season and return to a more normal weight in the off season. This weight fluctuation did not appear to increase rowers' prevalence of eating disorder cases, suggesting that while it is required for their sport, weight fluctuation did not increase their likelihood of having disordered eating. In addition, rowers did not differ from controls in their weight fluctuation, suggesting that runners may be experiencing a particularly low level of weight fluctuation. Body mass index was also different between the groups, but it would be expected that the runners would have a leaner physique due to the demands of the sport. This BMI is similar to that previously reported for runners. 31
This study failed to show an increased prevalence of disordered eating in collegiate distance runners versus controls. The reasons for this finding are unclear. Previous studies of runners 1,32,33 identified a range of 14–27.2% of runners scoring above concerning levels on the EAT instrument versus 11.3% probable case status in the current study. Consistent with the findings in the current study, Weight and Noakes 33 did not find an increase compared with controls in all but the elite runners. Smolak 24 similarly failed to find a significant association between eating disorders and running. It is interesting that these studies also found a higher prevalence of concerning behaviors in the elite versus collegiate population. This might suggest that the collegiate population is in general, compared with their elite counterparts, at a lower risk for eating disorders in athletes perhaps due to lower competitive expectations and a less selected population. Another possibility is that runners with disordered eating were missed in the current study. Given the dynamics of track meets it was not possible to ascertain what percentage of runners were missed, and it is quite possible that runners with eating problems avoided completing the questionnaire secondary to fear that coaches would learn of the results.
The finding of increased use of diuretics and laxatives by the rowers in this study is concerning. Sykora et al. 4 also found increased use of pathologic weight-making behaviors in lightweight rowers, but reported only increased use of vomiting. Another study 15 did not find an increased use of pathologic weight-making behaviors in lightweight rowers, who used food and fluid restriction only. While the increased use of laxatives and diuretics in the present sample of rowers was not associated with an increased prevalence of eating disorders, these behaviors are problematic. They may also be indicators of use of other pathologic and potentially harmful methods for weight loss, which were not investigated in this or previous studies. The use of pathologic weight-loss techniques, while they may not predispose athletes to development of eating disorders, can be an important clinical concern. Education and monitoring of these athletes for the use of pathologic weight-making behaviors needs to be further explored.
This study has several limitations. While the population of lightweight rowers was well represented, it was not possible to determine what percentage of distance runners at the meets actually completed the questionnaire. Thus the generalizability of the running sample, while seemingly unbiased, is in question. The weights and behaviors in this survey were assessed via self-report, and while participants were encouraged not to share answers or discuss the survey, some may have not answered truthfully. Research has shown that self-report weights are reasonably accurate. 34 In contrast, however, other studies 16 have indicated that conceptually complex behaviors such as binge eating are particularly difficult to assess via self-report. While the EDE-Q improves assessment of such behaviors by limiting the use of such words as binge eating and using operational definitions of behavior, some concepts are difficult to assess via self-report. Finally, participants may have been less than truthful due to concerns that the data might get back to their coaches and adversely impact their participation in their sport. Future directions might include verification of self-report answers with interview assessment.
The prevalence of eating disordered behavior that was found in this study does not appear to be different from the general collegiate population. These results do not seem to validate concerns that lightweight rowing for women at the collegiate level is associated with an increased prevalence of eating disorders, but do highlight the necessity to examine and address unhealthy weight management behaviors independently of eating disorder status. If that is the case, perhaps lightweight rowing for women may be successfully introduced to more colleges if it is closely tied to eating disorder and pathologic weight-making awareness, and proper selection of athletes naturally near the 130 pound weight limit.
Eating disorders in collegiate female athletes represent an important clinical concern for sports physicians, coaches, trainers, and administrators, all of whom have interest in protecting athletes from the potential health consequences of participation in sports. This study suggests that women can safely participate in lightweight rowing without undue concern that requiring them to make weight increases their prevalence of eating disorders. However, prevention of eating disorders and limitation of use of pathologic weight-making behaviors are ongoing clinical concerns. It is hoped that continued athlete education about safe weight making, and selection of athletes naturally close to the required weight of 130 pounds, will mitigate these concerns.
1. Wilmore J. Eating and weight disorders in the female athlete. Int J Sports Nutr 1991; 1:104–117.
2. Brownell K, Rodin J. Prevalence of eating disorders in athletes. In: Brownell K, Rodin J, Wilmore J, eds. Eating, Body Weight, and Performance in Athletes. Philadelphia: Lea & Febiger, 1992:128–143.
3. Fogelholm, M., Hiilloskorpi H. Weight and diet concerns in Finnish female and male athletes. Med Sci Sports Exer 1999; 31: 229–235.
4. Sykora C, Grilo C, Wilfley D, et al. Eating, weight, and dieting disturbances in male and female lightweight and heavyweight rowers. Int J Eating Disorders 1993; 14:203–211.
5. Sundgot-Borgen J. Risk and trigger factors for the development of eating disorders in female elite athletes. Med Sci Sports Exer 1994; 26:414–419.
6. Rucinski A. Relationship of body image and dietary intake of competitive ice skaters. J Am Dietetic Assoc 1989; 89:98–100.
7. Rosen L, Hough D. Pathogenic weight-control behaviors of female college gymnasts. Phys Sports Med 1986; 16:141–146.
8. Brooks-Gunn J, Warren M, Hamilton L. The relation of eating problems and amenorrhea in ballet dancers. Med Sci Sports Exer 1987; 19:41–44.
9. Brownell K, Steen S. Weight cycling in athletes: effects on behavior, physiology, and health. In: Brownell K, Rodin J, Wilmore J, eds. Eating, Body Weight, and Performance in Athletes. Philadelphia: Lea & Febiger, 1992;159–170.
10. Thiel A, Gottfried H, Hesse FW. Subclinical eating disorders in male athletes: a study of the low weight category in rowers and wrestlers. Acta Psychiatrica Scandinavia 1993; 88:259–265.
11. Dale KS, Landers DM. Weight control in wrestling: eating disorders or disordered eating? Med Sci Sports Exer 1999; 31:1382–1389.
12. Terry P, Waite J. Eating attitudes and body shape perceptions among elite rowers: effects of age, gender and weight category. Aus J Sci Med Sport 1996; 28:3–6.
13. Terry PC, Lane AM, Warren L. Eating attitudes, body shape perceptions and mood of elite rowers. J Sci Med Sport 1999; 2:67–77.
14. McCargar L, Simmons D, Craton N, et al. Physiological effects of weight cycling in female lightweight rowers. Can J App Physiol 1993; 18:291–303.
15. Morris F, Payne W. Seasonal variations in the body composition of lightweight rowers. Br J Sports Med 1996; 30:301–304.
16. Fairburn CG, Beglin SJ. Assessment of eating disorders: interview or self-report questionnaire? Int J Eating Disorders 1994; 16:363–370.
17. Wilson GT. Assessment of binge eating. In: Fairburn CG, Wilson GT, eds. Binge Eating: Nature, Assessment, and Treatment. New York: Guilford Press, 1993:227–249.
18. Black C, Wilson G. Assessment of eating disorders: interview versus questionnaire. Int J Eating Disorders 1996; 20:43–50.
19. Wilfley D, Schwartz M, Spurrell E, et al. Assessing the specific psychopathology of binge eating disorder patients: Interview or self-report? Behav Res Therapy 1997; 35:1151–1159.
20. Beglin S, Fairburn C. Evaluation of a new instrument for the detection of eating disorders in community samples. Psych Research 1992; 44:191–201.
21. Luce KH, Crowther JH. The reliability of the eating disorder examination—self-report questionnaire version (EDE-Q). Int J Eating Disorders 1999; 25:349–351.
22. American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edition.
(DSM IV). Washington DC: American Psychiatric Association, 1994.
23. Johnson C, Powers PS, Dick R. Athletes and eating disorders: the national collegiate athletic association study. In J Eat Disord 1999; 26:179–188.
24. Smolak L, Murnen SK, Ruble AE. Female athletes and eating problems: a meta-analysis. Int J Eat Disord 2000; 27:371–380.
25. Hoek HW. The distribution of eating disorders. In: Brownell KD, Fairburn CG, eds. Eating Disorders and Obesity: A Comprehensive Handbook. New York: Guilford Press, 1995:207–211.
26. Krahn DD, Kurth CL, Demitrack M, et al. The relationship of dieting severity to alcohol consumption in young women. J Subst Abuse 1992; 4:341–353.
27. Fairburn CG, Beglin SJ. Studies of the epidemiology of bulimia nervosa. Am J Psych 1990; 147:401–408.
28. Steiner H, Lock J. Anorexia nervosa and bulimia nervosa in children and adolescents: a review of the past 10 years. J Am Acad Child Adol Psych 1998; 37:352–359.
29. Killen JD. Weight concerns influence the development of eating disorders: a 4 year prospective study. J Consulting Clin Psych 1996; 64:936–940.
30. Fairburn CG, Cooper Z. The eating disorder examination. [12 ed.]. In: Wilson GT, Fairburn CG, eds. Binge Eating: Nature Treatment, and Assessment. New York: Guilford Press, 1993:317–332.
31. Slay H, Hayaki J, Napolitano M, et al. Motivations for running and eating attitudes in obligatory versus nonobligatory runners. Int J Eating Disorders 1998; 23:267–275.
32. Kiernan, M, Rodin J, Brownell K, et al. Relation of level of exercise, age, and weight-cycling history to weight and eating concerns in male and female runners. Health Psych 1992; 11:418–421.
33. Weight L, Noakes T. Is running an analog of anorexia? A survey of the incidence of eating disorders in female distance runners. Med Sci Sports Exer 1987; 19:213–217.
34. Stunkard AJ, Albaum JM. The accuracy of self-reported weights. Am J Clinical Nutr 1981; 34:1593–1599.
© 2001 Lippincott Williams & Wilkins, Inc.