Risk Factors for Disordered Eating in
High School and College Athletes
Mary Pritchard, Paul Rush, and BreeAnn Milligan
Boise State University
The present study compared disordered eating behavior in high school and college athletes. Results indicated significant differences between high school and college athletes in disordered eating, body dissatisfaction, and excessive exercise. Females were more likely to display disordered eating behaviors and body dissatisfaction than were males. Self esteem correlated with body dissatisfaction, disordered eating, and excessive exercise. Predictors of eating disordered behavior in college athletes were body dissatisfaction and self-esteem, with body dissatisfaction having more of an influence. For high school athletes, the primary predictor of disordered eating was body dissatisfaction followed by excessive exercise. Results from the present study suggests that when athletes are dissatisfied with their bodies, they are more likely to develop disordered eating and exercise behaviors. There are indications that coaches and trainers need to be aware of the possibility of disordered eating and exercise behaviors in athletes, and watch for the identified predictors of those behaviors.
According to the National Association of Anorexia Nervosa and Associated Disorders (2007), seven million women and one million men in the United States have a clinically diagnosable eating disorder. In a meta-analytic review article, Makino, Tsuboi, and Dennerstein (2004) reported that between 0.1% to 5.7% of females in Western countries have anorexia nervosa. The prevalence rates in Western countries for bulimia nervosa ranged from 0.3% to 7.3% in females and 0% to 2.1% in males.
Some researchers have suggested that the percentages of eating disorders in athletes may be greater than those exhibited in non-athlete populations. For example, in a collaborative study with the National College Athletic Association, Johnson, Powers, and Dick (1999) found that 9% of female collegiate Division I athletes needed treatment for their eating disorders. In addition, Johnson et al. reported that an additional 58% of female athletes met the criteria for exhibiting eating disordered behaviors (defined as eating and weight control behaviors, such as calorie restriction, body dissatisfaction, and excessive exercise, considered abnormal but not within the criteria for a formal diagnosis of an eating disorder; Anorexia Nervosa and Related Eating Disorders, Inc., 2005a, subclinical eating disorders section). Johnson et al. also found 1% of Division I male athletes had clinically diagnosable eating disorders and 38% exhibited disordered eating behaviors. In addition, a recent study of female high school athletes found that 20% exhibited disordered eating behavior such as eating restraint and vomiting during their competitive season (Pernick et al., 2006). There is also growing concern about the excessive exercise trends displayed by athletes (Anorexia Nervosa and Related Eating Disorders, Inc., 2005b; Bacon & Russell, 2004; Patel, Pratt, & Phillips, 2003). Research has also documented high levels of body dissatisfaction in athletes, especially in female athletes (Reinking & Alexander, 2005). Some researchers have suggested that athletes are more at risk for eating disorders and disordered eating behaviors if they participate in sports that require thinness (Patel, Pratt, & Greydanus, 2003; Smolak, Murnen, & Ruble, 2000). Other researchers have suggested that this may vary by gender (Milligan & Pritchard, 2006), with men in “lean” sports (e.g., wrestling, cross country) and women in “non-lean” sports (e.g., tennis, soccer) being more at risk than men in “non-lean” sports (e.g., basketball, tennis) and women in “lean” sports (e.g., gymnastics). Regardless of whether there is a difference in the type of sport or not, because athletes may be at risk for eating disordered behaviors and excessive exercise, it is essential to understand what factors influence athletes to engage in disordered eating and exercise behaviors. Only then will athletes, coaches, and athletic trainers have sufficient knowledge to predict and prevent disordered eating and exercise behaviors in athletes. Some of the factors that appear to predict disordered eating and exercise behaviors in athletes are discussed in the paragraphs that follow.
One factor that appears to influence athletes’ disordered eating and exercise behaviors is the pressure to perform their best. This pressure may come from external sources, such as coaches or parents, or from internal sources, such as the pressure athletes put on themselves to succeed. According to Thompson and Sherman (1999), coaches favor athletes who excessively exercise to increase performance, deny pain in order to keep training, conform with coaches/trainers requests, and pursue and accept nothing less than perfection. The pressures athletes place on themselves are generally attributed to personality and mentality to be the best regardless of the health risks (Kirk, Singh, & Getz, 2001). In fact, Hughes and Coakley (1991) argue that some athletes take their ‘sport ethic’ to the extreme, deviating into behaviors that have been expressly prohibited by their sport (e.g., steroid use). Western media also influences athletes’ expectations of the ideal body for their sport. A runner is tall and thin, a gymnast is short and petite, and a football player is muscular (Petrie & Rogers, 2001). Further, research has shown a correlation between exposure to the ideal body image and restrictive dieting and eating pathologies (Krane, Stiles-Shipley, Waldron, & Michalenok, 2001; Stice, Maxfield, & Wells, 2003; Thompson, Coovert, & Stormer, 1999). Athletes who do not fit the ideal body type for their sport are more likely to have both internal and external pressure to achieve that specific body image. For example, Brooks-Gunn, Burrow, and Warren (1988) suggested that when ability level is held constant, coaches often favor athletes with leaner body types. Sports such as gymnastics, cross country, and swimming demand a thin physique and low body weight/body fat. Bruch (1981) argued that some people with disordered eating view their behavior as a response to the demands put on them by others, not as something they want to do (see also Halliwell & Harvey, 2006; Hinton & Kubas, 2005).
Not only are the pressures and demands put on athletes factors that place them at an increased risk of developing eating disordered behavior, but studies have also found that a decrease in self-esteem, which can be caused by athletic pressures (Darnall, 2002), contributes to body image dissatisfaction (Abell & Richards, 1996; Gleason, Alexander, & Somers, 2000) and bulimic symptoms (Vohs, Bardone, Joiner, Abramson, & Heatherton, 1999). In a meta analysis, Marlowe, Schneider, and Nelson (1996) reported that individuals who fit the societal ideal of beauty are perceived to be more sociable, mentally healthy, and intelligent. People who believe that they meet this physical stereotypical standard will experience psychological benefits in their self-esteem, whereas those who do not believe they meet the stereotype will suffer a decrease in self esteem (Crocker & Knight, 2005). Further, low self-esteem has been shown to have a negative effect on dieting and bingeing behaviors in adolescents (Lindeman, 1994; Neumark-Sztainer, Beutler, & Palti, 1996). Friestad and Rise (2004) also found that having low self-esteem was a predictor for dieting, but it was only a significant predictor for males. Friestad and Rise also found that both body dissatisfaction and low self-esteem were predictive factors for eating disorders. Although athletes strive to meet an ideal body type, if this body shape is unattainable to that athlete, the athlete may engage in unhealthy thoughts and behaviors (Krane et al., 2001).
Picard (1999) has suggested that athletes who compete at higher levels of competition might be at an increased risk for developing an eating disorder. For an athlete to make it to the college-level, the athlete must be able to be able compete at a higher level of competition than when he or she was in high school. Athletes who want to be the best may put their health at risk by engaging in disordered eating and exercise behaviors in the hopes that they will improve their performance (Hughes & Coakley, 1991; Kirk et al., 2001). As might be expected, Picard (1999) found that female collegiate athletes at a higher level of competition were at a greater risk for disordered eating behavior than were athletes at a lower competition level.
Finally, there is some evidence that disordered eating, body dissatisfaction, and excessive exercise might influence the presence of one another. For example, Wichstrom (2000) found that body dissatisfaction and excessive exercise predicted disordered eating in Norwegian adolescents. Similarly, Heywood (2006) found that body dissatisfaction predicted disordered eating in college men and women. Adkins and Keel (2005) found that excessive exercise predicted disordered eating attitudes and behaviors in male and female undergraduates. Finally, Sim and Zeman (2005) found that body dissatisfaction predicted disordered eating in adolescent females. However, no studies have examined the relation between these variables in athletes.
Although many researchers have examined the influence of athletic status on the presence of disordered eating and exercise behaviors, no research has examined the impact of level of competition (high school v. college) on these behaviors in both male and female athletes. The goal of the present study was trifold. First, collegiate level competition is more competitive and pressure-driven than is high school level competition, especially in Division I collegiate athletics due to the levels of stress placed on collegiate athletes (Humphrey, Yow, & Bowden, 2000; Papanikolaou, Nikolaidis, Patsiaouras, & Alexopoulos, 2003). Thus, it was pertinent to compare the prevalence of disordered eating and exercise behaviors in high school and collegiate athletes. Second, research has established that, in general, female athletes are more at risk of body dissatisfaction, disordered eating and exercise behaviors than are male athletes (Johnson et al., 2004; Milligan & Pritchard, 2006). However, research on the influence of level of competition on disordered eating has focused on female athletes (Picard, 1999). Thus, less is known about the influence of level of competition on disordered eating and exercise in male athletes. Thus, the present study will examine the influence of level of competition on disordered eating and exercise behavior in both male and female athletes. Third, based on previous research, self-esteem, body dissatisfaction, and excessive exercise habits are known to relate to disordered eating. However, no studies have examined all of these factors simultaneously in athletes. It is important to examine how self-esteem, body dissatisfaction, and excessive exercise habits relate to disordered eating in athletes and whether those factors influence the two populations in the same way.
Based on previous research, it is hypothesized that college athletes will display more disordered eating, body dissatisfaction, and excessive exercise behavior than will high school athletes. Second, it is hypothesized that females will display more disordered eating behavior and body dissatisfaction than will males, whereas males will display more excessive exercise than will females. Third, it is hypothesized that individuals with a lower self-esteem will display more eating disordered, body dissatisfaction, and excessive exercise behavior than will individuals with higher self esteem. Finally, it is predicted that self esteem, body dissatisfaction, and excessive exercise will predict disordered eating in both groups.
A total of 354 athletes residing in a northwestern community participated in this study. The sample population of athletes consisted of 139 females and 215 males. The age of the participants ranged 14 to 23 (M = 18.09, SD = 2.13).
For the purpose of the present study, the sample population was broken down into high school and college athletes. A total of 176 Division I student-athletes from an institution in the Western Athletic Conference participated in this study. The sample population consisted of 99 females (56%) and 77 males (44%) from a variety of sports (e.g., soccer, track, baseball, wrestling, golf, basketball, gymnastics, skiing, tennis). The age of the participants ranged from 18 to 23 years with a mean age of 19.84 (SD = 1.44). Due to the limited racial diversity at the institution surveyed, the Institutional Review Board would not allow us to ask specific racial classifications so as to not be able to identify any athlete’s specific responses. However, based on the racial make up of athletes at this university, we estimated that 90% of athletes were Caucasian.
In order to control for differences in region of the country or state, we felt it was important to recruit high school athletes from the same geographic area as the collegiate athletes. A total of 178 varsity and junior varsity athletes from local high school participated in this study. The sample population consisted of 17 female junior varsity athletes, 23 female varsity athletes, 42 male junior varsity athletes, and 96 male varsity athletes from a variety of sports (e.g., swimming, football, soccer, track, baseball, wrestling, golf, softball, basketball, volleyball, cheerleading; there were no differences in disordered eating between varsity and junior varsity players). The age of the high school participants ranged from 14 to 18, with a mean age of 16.37 (SD = 1.00). Once again, due to the limited racial diversity in the community, the Institutional Review Board would not allow us to ask specific racial classifications so as to not be able to identify any athlete’s specific responses. However, based on the racial make up of athletes at the schools that completed the survey, we estimated that 90% of athletes were Caucasian.
Body image dissatisfaction. Body image was assessed by the 12-item Body Shape Questionnaire (Cooper, Taylor, Cooper, & Fairburn, 1987; see Cooper et al. for discussion on validity). The survey asked about athletes’ perceptions of their bodies (e.g., Have you pinched areas of your body to see how much fat there is?). Responses were rated on a 6-point scale (1=never, 6=always) and were averaged. Thus, scores could range from 1 to 6, with higher average scores indicating greater body dissatisfaction. This questionnaire demonstrated good reliability in the present study (Chronbach’s α = .88).
Self-esteem. Levels of self-esteem were measured using the 9-item Rosenberg Self-esteem scale. This scale has been shown to be both valid and reliable (Chronbach’s α = .79; Rosenberg, 1989). This scale asks individuals to assess their feelings about themselves (e.g., "I feel I have a number of good qualities."). Responses were measured on a 4-point scale (1=strongly disagree, 4=strongly agree) and a mean score was calculated. Thus, scores could range from 1 to 4, with higher scores indicating greater self esteem.
Disordered eating behaviors. Disordered eating behaviors were assessed by the Eating Attitude Test (EAT-26; Garner, 1993; 1997). The EAT-26 consists of 26 forced-choice items measured on a 6-point scale, where items marked never, rarely, or sometimes are scored 0, often is scored 1, usually is scored 2, and always is scored 3. The EAT-26 has questions relating to dieting behaviors, bulimia behaviors, preoccupation with food, and oral control (e.g., “I avoid eating when I am hungry”). In addition to utilizing raw summation scores on the EAT-26, the EAT-26 uses a cutoff score of 20 to determine if a person is at risk for a clinical eating disorder. Thus athletes were classified as at risk or not at risk for an eating disorder based on their cutoff score. The EAT-26 was chosen for this study because it has been shown to be reliable and valid in several populations (Mintz & O’Halloran, 2000), including athletes (Kirk et al., 2001; Picard, 1999) and demonstrated adequate reliability in this study (Chronbach’s α = .94).
Exercise behaviors. Excessive exercise was assessed by the Exercise Dependence Questionnaire (Hausenblas & Downs, 2002). The Exercise Dependence Questionnaire consists of 29-items (e.g., I feel irritable if I cannot exercise) measured on a 5-point Likert scale, where 1: never, 5: always. Individual item responses were averaged to create a score. See Hausenblas and Downs for reliability and validity data. This questionnaire demonstrated good reliability in the present study (Chronbach’s α = .94).
Prior to initiating the study, the Institutional Review Board approved all procedures both high school and collegiate athletes. For the collegiate athletes, researchers coordinated with the coaches of each sport to survey teams at a convenient time. Athletes were assured that their participation was voluntary and confidential. As the survey was anonymous, consent for collegiate athletes was implied with completion of the survey. If the athlete did not wish to disclose information on the survey, he/she was told they could skip any item or discontinue participation at any time.
Due to the age of the high school participants (under age 18) and the sensitive nature of the subject, extra care was taken to ensure that all parties involved were familiar with the nature of the survey. First, researchers first obtained consent from the Boise School District Review Board. Second, principals at the four high schools in the district were contacted. Third, if the principals approved, two times, either before or after practice, were set up to hand out the parental consent form. Finally, with the parent’s approval the athletes were invited to participate in the research. Each athlete was then asked to sign an assent form. If the athlete assented, he or she was administered the survey. Participation was completely voluntary and confidential. All athletes, high school and collegiate, were informed they did not have to fill out any sections that made them uncomfortable.
The present study had four main hypotheses: 1) college athletes will display more disordered eating, body dissatisfaction, and excessive exercise behavior than will high school athletes. 2) females will display more disordered eating behavior and body dissatisfaction than will males, whereas males will display more excessive exercise than will females. 3) athletes with a lower self-esteem will display more eating disordered, body dissatisfaction, and excessive exercise behavior than will athletes with higher self esteem. 4) self esteem, body dissatisfaction, and excessive exercise will predict disordered eating in both groups. Each of the four hypotheses is addressed in its own section below.
The Influence of Competition Level (College v. High School) on Disordered Eating Behavior, Body Dissatisfaction, and Excessive Exercise
To test the hypothesis that college athletes would display more disordered eating, body dissatisfaction, and excessive exercise behavior than would high school athletes, independent samples t-tests were conducted. Before examining whether level of competition influenced, we wanted to see what percentage of the athletes had scores above the cutoff score of 20 on the EAT-26, the score used to determine if a person is at risk for a clinical eating disorder. In the present study, 8.4% of high school athletes had a score of 20 or higher on the EAT 26, whereas 28.2% of college athletes had a score of 20 or higher. Not surprisingly, there was also an effect of competition level on disordered eating behavior, t (349) = 6.34, p < .001, with athletes in college being at a greater risk of developing eating disorders than those in high school athletics (see Table 1 for means and standard deviations). As predicted, there was a significant effect of competition level on body dissatisfaction, t (349) = 4.68, p < .001. Athletes in college displayed more body dissatisfaction (M = 2.52, meaning collegiate athletes were somewhere between fairly satisfied and neutral about their bodies) than did those in high school (M = 2.00, meaning high school athletes were fairly satisfied with their bodies; see Table 1 for means and standard deviations). Finally, there was an effect of competition level on excessive exercise, t (351) = 3.07, p < .01. College athletes displayed more excessive exercise (M = 2.78, indicating that they often exercised excessively) than high school athletes (M = 2.54, indicating that high school students excessively exercised somewhere between sometimes and often).
The Influence of Gender on Disordered Eating, Body Dissatisfaction, and Excessive Exercise
To examine the influence of gender on disordered eating, body dissatisfaction, and excessive exercise, we conducted independent samples t-tests. Before examining whether gender influenced disordered eating behavior, we wanted to see what percentage of males and females had scores above the cutoff score of 20 on the EAT-26. In the present study, 33.1% of women had a score of 20 or higher on the EAT 26, whereas 8.4% of men had a score of 20 or higher. Not surprisingly and as predicted, there was an effect of gender on the amount of disordered eating behaviors displayed, t (349) = 8.19, p < .001, with females being at a greater risk than males (see Table 1 for means and standard deviations). There was also a significant effect of gender on body dissatisfaction, t (349) = 11.50, p < .001, with females displaying greater levels of body dissatisfaction (M = 2.98, indicating that most females were relatively neutral about their bodies) than males (M = 1.80, indicating men were fairly satisfied with their bodies; see Table 1 for means and standard deviations). However, contrary to our hypothesis, there was not a significant influence of gender on excessive exercise, t (351) = .34 (see Table 1 for means and standard deviations).