EXPLORING THE SOCIOCULTURAL PATHWAY MODEL AS IT PREDICTS BODY SATISFACION IN ATHLETES AND NONATHLETES

This study was a secondary data analysis in which the sociocultural pathway model (Stice & Agras, 1998) was tested in male and female athletes and nonathletes. There were three primary purposes for this study. The first aim was to test the applicability of the sociocultural model theory in the current sample by conducting confirmatory factor analyses on the measurement constructs. The second aim was to create a structural equation model that fit the data best according to theoretical directionality of the constructs in the sociocultural theory. The third aim was to compare the predictive paths of this model across sexes and athlete subgroups. The sample (N=543) was composed of competitive athletes (n=56), recreational athletes (n=192), and non-athletes (n=294) who completed an online questionnaire measuring the major components of the sociocultural pathway model. Construct means were examined in nonathletes and athletes of both sexes using two-way multivariate analysis of variance, follow-up two-way analyses of variance, and post-hoc Tukey HSD tests. Confirmatory factor analyses were run on the original factor structures of each construct and found model fit to be poor. Measures were revised, cutting poorly loading items and confirmatory factor analyses were conducted on revised measures, finding model fit to be adequate. A structural equation model predicting body satisfaction from the internalization of thin ideals, negative affect, positive affect, and perfectionism had low-adequate fit in the sample. Comparisons of the predictive model revealed that the model fit to be adequate across group. Findings suggest the field is in need of improved operational definitions and more psychometrically sound measures of body satisfaction. Attention to measurement of other constructs of the sociocultural pathway model is also warranted. Future research should examine measurement structures and invariance across a range of minority groups, as well as longitudinal models to test and expand the applicability of this model of eating disorders prevention.

. Historically, the field of has found ample evidence to support that female athletes, performing in lean, aesthetic sports, are at risk for developing subclinical eating disorder symptoms, defined as those that do not meet current DSM 5 criteria for a clinical eating disorder, yet are worrisome eating problems and body image or weight concerns (Beals & Manore, 1994;Sanford-Martens et al., 2005). Since Rosen et al. (1986) found that a substantial percentage of female athletes reported pathogenic weight controlling behavior, some researchers have found support for more eating pathologies in female athletes versus non-athletes Sundgot-Borgen & Torstveit, 2004;).
However, there are methodological and theoretical shortcomings to the research literature as a whole. These methodological shortcomings are primarily explained by the limitation of instruments used by the field. Originally constructed to measure symptoms in athletes, the Eating Disorder Inventory (Garner, Olmstead, & Polivy), and the Eating Attitudes Test (Garner & Garfinkel, 1979) have been used inappropriately to find prevalence rates in athletes, a practice which has been at least partly to blame for inconsistent findings in the field (Hausenblas & McNally, 2004). Prevalence rates based solely on these scales are inappropriate.

Media's Impact on the Development of Eating Disorders in America
The role of the media in the development of eating disorders is unequivocal. As it is the most consumed avenue delivering messages containing social standards, media is a powerful force in upholding the status quo. Media has a particularly strong role in the most empirically supported model of eating disorder development, the sociocultural model (Groesz, Levine, & Murnen, 2002;Thompson & Stice, 2001). One of the most significant variables leading to eating disorder development, according to the sociocultural model, is the internalization of thin ideals (Thompson & Stice, 2001).
These ideals are symbolized in consumed forms via television series, commercials, print ads, movies, music, and recently, social media (Mabe, Forney, & Keel, 2014). Several researchers have found links between media exposure and body dissatisfaction and eating disorder symptomatology (Harrison & Cantor, 1997;Stice et al., 1994). In a study by Nichter and Nichter (1991), adolescent girls described the ideal girl as 5 ft. 7 in., 100 lb., size 5, with long blond hair and blues eyes. This depiction supports that, from an early age, both American boys and girls are inundated with media images and messages of what it means to be beautiful. For girls, this ideal image is of a thin, lean woman; while for boys, this ideal image is of a strong, muscular man. Psychologists argue that it is not the images themselves that might lead to eating disorders, but the extent to which individuals internalize their underlying messages, such as "I must be thin and beautiful to succeed", that might lead to an eating disorder. Still, this model fails to explain why, if most members of the population receive these messages through the media, only a small number actually go on to develop an eating disorder. It is also important to note that other institutions such as family, peers, schools, athletics, and businesses support this idealized image of beauty (Thompson & Stice, 2001) to varying degrees.
Feminists describe these media messages and images as explicit forms of sexual objectification of the female body (Tiggemann, 2013). This theory is known as objectification theory and posits that the female body is socially constructed as an object meant to be evaluated by its appearance. It is suggested by Fredrickson and Roberts (1997) that a woman's repeated experiences of objectification will eventually lead her to arrive at the conclusion of the observer's, usually a man's, appraisal of her body. That is to say a woman begins to evaluate herself, or survey her body as an object based on appearance, a process termed "self-objectification" (Fredrickson & Roberts, 1997). Self-objectification can lead to many behavioral and cognitive responses, including shame, anxiety and guilt when a woman's body does not meet those social standards shown in the media. Among the behavioral consequences of selfobjectification are eating disorders (Tiggeman, 2013). Feminist theory recognizes that, although all Western women do live in a culture that sexually objectifies women, not all women will internalize the observer's evaluation of their bodies. Self-objectification is also seen as a trait that is stable over time, which some women, perhaps who go on to develop an eating disorder, possess (Slater & Tiggemann, 2012). Regarding the link between self-objectification and eating disorders, Calogero (2009) and Tylka & Sabik (2010) found support for the mediating relationship between body shame, selfobjectification, and disordered eating in community samples. Moradi, Dirks, and Matterson (2005) provide additional support for the inclusion of self-objectification in models of eating disorder development as they found that in their sample of N=221 undergraduate women, sexual objectification experiences, the internalization of those experiences, body surveillance, and body shame were all significant predictors of eating disorder symptomatology.
Self-objectification theory provides other possible explanations for why not all women develop eating disorders despite being consumers of the same social standards and ideals. A study by Harper and Tiggemann (2008) sought to test the variable of state self-objectification, meaning the varied responses one gives contingent on the environment. By giving participants a set of thin idealized images of females, Harper and Tiggemann (2008) ascertained that state self-objectification, weight-related appearance anxiety, body dissatisfaction, and negative mood were elicited in some female participants. However, not all participants reported state self-objectification, giving credence to the theory's postulate that not all women will self-objectify; thus, not all women will go on to show behaviors, such as eating disorders, that result from selfobjectification.
Pioneering neuropsychology research also suggests self-objectification theory presents a plausible biological explanation for why not all women internalize ideals of beauty and go on to develop eating disorders (Riva et al., 2014). In this analysis, selfobjectification is considered a cognitive process in which a woman uses an allocentric frame of reference to remember times in which she evaluates herself based on her appearance (Riva et al., 2014). The allocentric ("from outside") lock hypothesis suggests that women with eating disorders view themselves from the outside.
Interestingly, this hypothesis views the source of the distorted body image in the cognitive processes of the individual, positing that the woman is locked into a negative memory of her body (from repetitive negative objectifications) and is no longer sensitive to bodily changes (Riva et al., 2014). These theories identify cognitive processes as the reason why only some and not all Western women develop an eating disorder, despite all receiving messages of the thin ideal.
This discussion of self-objectification theory leads to further description of other theoretical models that have been developed to explain the development of eating disorders.

Theoretical Models
The sociocultural model is the most widely accepted and scientifically scrutinized theoretical model explaining eating disorder development. Although the field has historically accepted this model, there are several other theoretical models that also seek to explain eating disorder etiology. These models include cognitive behavioral theory, feminist theory, and neuropsychological explanations for eating disorder development. Noting that feminist theory and neuropsychological models have been discussed, further explanation of cognitive behavioral theory is included in this section. Additionally, support for the sociocultural model of eating disorder development is reviewed.
Behaviorists originally thought eating disorders developed out of fears of fatness and body image disturbances (Russell, 1979). Over time Behavioral theory integrated concepts from cognitive psychology, giving the field a cognitive-behavioral model of eating disorder development. This model includes: a) the body self-schema, b) cognitive biases, c) binge eating, d) compensatory behavior, e) negative reinforcement of compensatory behavior by reduction of negative emotion, and f) psychological risk factors (Williamson, White, York-Crowe, & Stewart, 2004). Several studies support this cognitive behavioral model including Williamson et al. (1999) in which the researchers proposed that individuals with anorexia nervosa and bulimia nervosa develop schemas related to body and weight related information. These schemas influence the reactions of those with eating disorders, such as binge eating, purging, or self-starvation (Reas, Wisenhunt, Netemeyer, & Williamson, 2002). It is then hypothesized that body and shape concerns impair processing, which lead to the maintenance of the eating disorder. Some cognitive biases that result are attention bias, selective memory bias, selective interpretational bias, body size overestimation, and extreme drive for thinness (Williamson et al., 2004). Thus, it is the goal of cognitive behavioral therapy to bring these flawed cognitive processes to the forefront of the individual's attention so that they can be restructured (Williamson et al., 2004). At least three prevention programs have been developed which utilize the cognitive behavioral model of eating disorder development and treatment.
Stice and Agras (1998) developed a theory suggesting that the internalization of thin ideals leads to body dissatisfaction which in turn leads to negative affect and dieting behaviors which increase the risk for eating disorder symptomatology, specifically bulimic pathology. This theory has been labeled the sociocultural model of eating disorder development. Support for the model is expansive (Bradford & Petrie, 2008;Brannan & Petrie, 2008;Leung, Geller, & Katzman, 1996;Mintz & Wright, 1993;Ricciardelli & McCabe, 2004;Striegel-Moore & Cachelin, 1999 (Suisman et al., 2012;Suisman et al., 2014), maternal characteristics (Le Grange et al., 2014), interpersonal variables (Goldschmidt et al., 2014), bullying (Vartanian et al., 2014), trauma (Groleau et al., 2012;Sachs-Ericsson et al., 2012), and emotion dysregulation (Lavendar & Anderson, 2010). With research supporting the integration of other variables into the sociocultural model, the field is burdened with the responsibility of deciding which factors and variables place individuals at risk for the onset of an eating disorder. Having inconsistent models explaining the same phenomena makes it difficult to create prevention programs that have the correct aims and goals for minimizing the risk for the onset of eating disorders. Nonetheless, some prevention efforts have been taken in the field, with sparing utility.

Prevention Programs
In a meta-analysis of eating disorder prevention programs, Stice, Shaw, and Marti (2007) found that 51% of programs (N=51) reduced eating disorder risk factors and 29% reduced current or future eating pathology. Stice and Shaw (2004) also found that of the 38 eating disorder prevention programs evaluated, only six produced reductions in current or future eating psychopathology symptoms, and only two produced effects that were replicated. Despite these disappointing statistics, several prevention programs have been found to empirically reduce eating disorder risk and symptomatology (Stice, Shaw, & Marti, 2007). Furthermore, the most effective prevention programs have foundations in theoretical models of eating disorder development. These programs include Girl Talk (McVey et al., 2003), Student Bodies (Taylor et al., 2006), Body Project (Stice et al., 2006), and Healthy Weight (Stice et al., 2006).
These four successful prevention programs have several factors in common.
First, each of these programs identifies young girls, ages 15-19, as the most efficient target group to receive the program (McVey et al., 2003;Stice et al., 2006;Taylor et al., 2006). Developmental psychopathology supports targeting this age group in prevention efforts since the onset for eating disorders occurs during the adolescent years (Striegel- Moore et al., 2003). Second, each of these programs targets women, since females are at significantly higher risk for the development of eating disorders than males (Newman et al., 1996). Finally, the goals of the programs were to reduce eating pathology by increasing resistance to sociocultural pressures of an ideal body, and to strengthen body satisfaction and self-esteem (Stice, Shaw, & Marti, 2007). A summary of each of the programs can be found below.
Girl Talk is a school-based prevention program, which seeks to improve body esteem and global self-esteem through the use of peer support groups (McVey, Lieberman, Voorberg, Wardrope, & Blacmore, 2003). Research suggests that young girls are impressionable and influenced by the attitudes of their peer groups (Lieberman, Gauvin, Bukowski, & White, 2001). For this reason, Girl Talk hopes to deliver a group that will improve body and self-esteem. The group is designed to meet for 10-sessions and is delivered by nurses trained on the manualized program, Every BODY Is A Somebody (EBIS) (Seaver, McVey, Fullerton, & Stratton, 1997). EBIS was designed to teach young girls the dangers of the ideals of beauty through media literacy and to promote the importance of self-esteem. With the emphasis on media literacy to reduce the internalization of an ideal image of beauty, Girl Talk has theoretical grounding in the sociocultural model of eating disorder development. Initial findings by McVey et al. (2003) were encouraging as they showed the group led to increases in weight-related esteem and decreases in dieting and these effects remained stable at 3-month follow-up.
These results are promising for the effectiveness of peer groups to decrease risk factors for eating disorder development.
Student Bodies is an 8-week, Internet based prevention program designed to prevent the onset of eating disorders in at risk populations of young women (Taylor et al., 2006). Rooted in cognitive-behavioral theory, the intervention's goals are two-fold, to reduce: 1) body dissatisfaction and 2) excessive weight concerns (Taylor et al., 2006). Program design requires participants to participate in an online discussion group, monitored by a clinical psychologist, in which they are prompted to read content and then complete assignments. This is similar to the cognitive-behavioral intervention, self-monitoring. The first results for Student Bodies as an effective prevention program were encouraging. Most important was that a brief, 8-week, Internet-based intervention led to reductions in weight and shape concerns in at risk, young women, which were sustained through 12-month follow-up (Taylor et al., 2006). Additionally, Taylor et al. (2006) found that a reduction in weight and shape concerns was associated with a reduction in the onset of eating disorders, a potential major finding supporting the causal relationship between weight and shape concerns and eating disorder development.
Body Project and Healthy Weight (Stice, Shaw, Burton, & Wade, 2006) were the two programs that had effects replicated across labs, as found by Stice and Shaw (2004). Body Project is a dissonance-based program, in which the ideal image of beauty is challenged through verbal, written, and behavioral exercises. The target population for this program is at-risk girls with body image concerns (Stice, Presnell, Gau, & Shaw, 2007 (Stice et al., 2008), when high school staff delivers the program (Stice et al., 2009;Stice et al., 2011), and when it is presented online (Stice et al., 2012).
Healthy Weight's focus is on the promotion of lasting healthy changes to eating and exercise as the means to maintaining a healthy body weight and satisfaction in atrisk girls (Stice, Presnell, Gau, & Shaw, 2007). Participants create an individual lifestyle change plan with the support from a group facilitator and other group members.
A key component of the Healthy Weight intervention is the use of motivational interviewing. Previous trials of Healthy Weight have found that it is associated with sustained reductions in body dissatisfaction, dieting, negative affect, and bulimic symptoms Stice et al., 2003).
Despite these promising results, however, eating disorders prevention programs still need to improve their efficacy by targeting the root causes and risks for eating disorders (Bulik, 2013). Having inconsistent models of eating disorder development makes it difficult to create prevention programs that accurately intervene on the most important risk variables. Additionally, the effective prevention programs discussed in this report are each targeting young, at-risk girls. Although prevalence rates are lower, young boys and minority groups increasingly also experience clinical and subclinical eating disorders. As such, prevention programs need to work to include young boys and minorities better.

The Inclusion of Minorities
Perhaps one explanation for the small representation of minorities in the field of eating disorder epidemiology, research, treatment, and prevention is that eating disorders go undiagnosed in these populations. Strictly adhering to DSM criteria could be limiting for many clinicians who serve minority populations, including men. For example, many men present with body image concerns and disordered eating behaviors but have normal weights and appear healthy. If a clinician expects to see a man present with the same diagnostic criteria as a woman, he or she might miss a clinically relevant eating disorder (Cain et al., 2012). Additionally, many minority populations do not seek treatment for disordered eating, making it difficult to study eating disorders in these populations (Becker et al., 2010;Marques et al., 2011).
Given that prevention efforts are primarily grounded in sociocultural explanations for eating disorder development, it is important for the field to gain a clearer understanding of how everyone, including minorities internalize thin ideals.
Research by Stark-Wroblewski, Yanico, and Lupe (2005) on the impact of acculturation in Japanese and Chinese international student women reveals that individuals from countries once thought to be void of the Western ideal of beauty begin to internalize thin ideals after a period of time in the United States. However, is this true for women from other countries across the globe who immigrate to America? Also important would be the exploration of how eating disorder behaviors and attitudes manifest in minority populations. For example, Latino women are more likely to binge and purge rather than restrict their food intake because of cultural factors (Franko, 2007). Gaining a broader understanding of what behaviors are more prevalent and what cultural factors might protect ethnic minorities (Warren et al., 2005) from developing an eating disorder could aid in enhancement and tailoring of prevention programs.
There also is a growing body of research devoted to eating disorders in the homosexual population. Several research studies have found higher rates of eating disorders in homosexual men than in heterosexual men (French et al., 1994;Russell & Keel, 2002). In contrast, research is still unclear as to whether or not homosexual women are susceptible to the same standards of beauty as heterosexual women; thus, are protected from eating disorder development (Barron, 1998;Brown, 1987;Lakkis, about eating disorder development in the homosexual population is a concern and needs to be addressed so that preventive efforts can be inclusive. Despite the growing research on eating disorders in men, little is being done by prevention programs to include men (Cain et al., 2012). For example, Cain et al. (2012) found that the onset of eating, weight, and shape concerns in men occurs during college.
Considering prevention efforts are primarily targeted at adolescent girls, this finding encourages researchers to develop prevention programs for college-aged men.
Furthermore, taking into account that weight control behaviors have increased in men over the past two decades (Chao et al., 2008), it is important to recognize gender bias as it exists in the field and to begin to include boys and men in research and prevention efforts.

Risk Factors for Athletes
Athletes are thought to be at risk for developing body image dissatisfaction and subclinical eating disorder symptoms ). An athlete's performance rests primarily on the ability of their body to perform at an optimal level. This pressure to perform and succeed is both internal and external, with pressures coming from the self and from coaches, peers, family, and society (Beals & Manore, 1994). Given that athletes also receive societal pressures to adhere to an ideal image of beauty, it is thought that with the addition of sport specific pressure, athletes are more at risk for body image dissatisfaction and subclinical eating disorder symptoms (Thompson & Sherman, 2010). Sport environments that stress the importance of the body in competition shift the athlete's focus on performance to the body (Anderson, Petrie, & Neumann, 2012;Coppola, Ward, & Freysinger, 2014;Kerr, Berman, & De Souza, 2006;Muscat & Long, 2008). Thus, studies have found that athletes are more willing to engage in weight-controlling behaviors and become dissatisfied with their bodies when they fail in competition (Sudgot-Borgen, 1994;Thompson & Sherman, 2010 Thompson & Sherman, 1999). Also, competitive thinness may result in an athlete's dissatisfaction with his/her body when he/she is compared to another athlete (Arthur-Cameselle & Baltzell, 2012).
Furthermore, elite athletes are also motivated by external factors such as awards and monetary compensation for their success (Anderson, Petrie, & Neumann, 2011;Goldfield, 2009). This added layer of pressure could be an explanation for why some research has found that elite athletes suffer from disordered eating more than recreational athletes, as they are willing to go to greater lengths, such as weight compensatory behaviors, to ensure their bodies perform at the highest level (Kong & Harris, 2014;Torstveit, Rosenvinge, & Sundgot-Borgen, 2008). These behaviors might exist more in sports that require speed or endurance (i.e. "the lighter I am, the faster and further I will go"), scores/results given on the basis of appearance (i.e. "the thinner I look in my uniform, the higher scores I will receive"), or are weight dependent (i.e. "I must lose/gain five pounds to meet my weight class). However, whether or not competition level increases an athlete's risk for developing body image dissatisfaction or subclinical eating disorder symptoms is unclear since some studies have found fewer eating pathologies in elite athletes (Hausenblas & Carron, 1999;Hausenblas & McNally, 2004 1990). These characteristics may place athletes at risk for eating disorder symptoms.

Disorder Symptoms
Disordered eating behaviors and attitudes primarily affect women (Hudson, Hiripi, Pope, & Kessler, 2007). According to Espíndola and Blay (2006), 95% of disordered eating cases occur among women. Research by Greenleaf et al. (2009) revealed that in a sample of N=204 college athletes, female athletes suffered from disordered eating at a higher rate than male athletes, with 25.5% (n=52, N=204) reporting as symptomatic. In comparing male and female athletes, Engel et al. (2003) found that women had higher drive for thinness and body dissatisfaction scores, and reported more purging and restriction behaviors than men.
For this reason, most research has focused on disordered eating in women.
Furthermore, female athletes are at heightened risk for developing disordered eating behaviors and attitudes as they are also susceptible to Western ideals of beauty for women Kong & Harris, 2014). Therefore, female athletes rather than male athletes participating in lean, endurance, and aesthetic sports are thought to be at higher risk for developing disordered eating behavior and attitudes (Hausenblas & Carron, 1999).
However, male athletes do suffer from subclinical eating disorder symptoms, although at a lesser rate than female athletes (Hausenblas & Carron, 1999;Hausenblas & McNally, 2004) and do seem to also suffer from a gender bias in the field (Thompson & Sherman, in press). Given the potential severity of symptomatic eating disorder behaviors (i.e. negative affect, body image concerns) and the possibility they go undetected, it is dangerous and unethical to exclude male athletes in studies examining eating disorder symptoms. In fact, Sanford-Martens et al. (2007) found more male than female athletes had subclinical eating disorder symptoms. Body image concerns are also evident in male athletes (MacKinnon et al., 2003) and should be treated as seriously as those of female athletes as they could potentially lead to subclinical eating disorder symptoms (Stice, 2001). Additionally, models explaining eating disorder symptom development need to be refined and tested as they show poor fit in male athletes (Petrie et al, 2013).

Disorder Symptoms
Sport type may be important in determining an athletes' level of risk for body image dissatisfaction or subclinical eating disorder symptoms. In fact, Engel et al. (2003) found that sport type accounted for the most variance in eating disorder indicators in a sample of 1445 elite Division I athletes. Research suggests that athletes participating in lean sports that stress a thin body build, such as gymnastics, may have higher incidence of body image dissatisfaction and subclinical eating disorder symptoms than athletes participating in sports allowing normal builds, such as ball sports (Beals & Manore, 1994;Kong & Harris, 2014;Krentz & Warschburger, 2011;Swami, Steadman, & Tovée, 2009 Interestingly, however, this same sample of elite gymnasts identified weight-related performance pressures (e.g. "thin is going to win") rather than body image dissatisfaction as reasons for dieting (de Bruin, Oudejans, & Bakker, 2007). This finding suggests dieting and body image present differently in aesthetic sports than in non-lean sports. Disordered eating behaviors and attitudes have also been found in other lean sports including wrestling and gymnastics (Engel et al., 2003), synchronized swimming (Ferrand et al., 2009), dance, and gymnastics (Francisco, Alarcão, & Narciso, 2012), Conversely, results from a prevalence study conducted by Greenleaf, Petrie, Carter, and Reel (2009) did not find that eating disorder symptoms were more prominent in aesthetic, lean sports. Hausenblas and Carron (2002) proposed classifying team sports into five groups: endurance (cross country, track, swimming), aesthetic (cheerleading, diving, gymnastics, synchronized swimming), power sports (downhill skiing, crew/rowing), ball game (basketball, softball, soccer, volleyball, field hockey, lacrosse) and technical (tennis, golf). In a sample of N=204 female, collegiate athletes, separated into these five classifications, there was no significant association between sport type and eating disorder status ).
Likewise, Hausenblas and McNally (2004) found that in a sample of n=217 track and field athletes, there were no sport-group differences for either eating disorder symptoms or prevalence in middle/long distance, sprinters, or field athletes. Other studies have also found that body image dissatisfaction and subclinical eating disorder symptoms did not vary by sport type (Karr et al., 2012;Kong & Harris, 2014;Voelker, Gould, & Reel, in press).

Sports as a Potential Protective Factor
Some of the research on body image dissatisfaction and subclinical eating disorder symptoms suggests that athletes are at greater risk for developing disordered eating behaviors and attitudes. Conversely, there is evidence for sports offering some protection from disordered eating behaviors and attitudes in athletes. It is possible that athletes are protected from negative affect due to the physiologic and psychological effects of exercise (Gaspar et al., 2011;Hausenblas & Fallon, 2006;Hausenblas & McNally, 2004) and that athletes actually have fewer disordered eating attitudes (e.g. body image) (Hausenblas & Downs, 2001) as their bodies fall more in line with societal standards of beauty (i.e., leanness) (Cook, Hausenblas, Tuccitto, & Giacobbi, 2011). In one study (Hulley, Currie, Njenga, & Hill,2007), Kenyan runners (n=75) were influenced by the physiological and psychological benefits of exercise as they reported the least eating disorder psychopathology as compared to controls and UK runners.
However, this study also includes a cultural confound since the UK runners, influenced by a Western ideal of beauty, reported significantly higher levels of psychopathology and eating pathologies (Hulley, Currie, Njenga, & Hill, 2007).
Similar protective factors of sport have been found in a German study of Elite high school athletes (n=576) compared to a control group (n=291) of non-athletes (Rosendahl et al., 2009). There were no differences between athletes and non-athletes on levels of disordered eating. The researchers suggest that coach involvement might be another reason for this protective effect of sports, also supported by Francisco, Alarcão, and Narciso (2012), but they did not suggest any other plausible explanation for how sports may have been protective in their sample (Rosendahl et al., 2009).
Other studies do not support the claim that athletes are more at risk for developing subclinical eating disorder symptoms (Davis & Cowles, 1989;de Oliveira Coelho, Soares, & Ribeiro, 2010;DiBartolo & Shaffer, 2002), and, instead, find that athletes possess higher levels of body image satisfaction and less disordered eating symptoms than nonathletes (Hausenblas & McNally, 2004).
Furthermore, Estanol, Shepher and MacDonald (2013) support the idea that the mental processes of athletes might protect them from developing negative pathologies, such as eating disorders. In their sample of 205 female ballet dancers, Estanol, Shepherd and MacDonald (2013) found that mental skills, including self-confidence, weakened the relationship between negative affect and disordered eating. Thus, having access to a sport psychology consultant or a coach who teaches self-confidence may be a protective factor from body image dissatisfaction and/or eating disorder symptoms.
This study has implications for the efficacy of mental skills as effective coping mechanisms that could be taught in prevention programs aimed at athletes.

Rationale for Current Study
Body image disturbances are of rising public health concern. Defined as the internal representation of one's outer appearance (Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999), having a poor body image can lead to psychological, psychosocial, and medical complications (Thompson et al., 1999). The most widely accepted and researched results of body image problems are clinical and subclinical eating disorders. Models have included sociocultural pressures, personal variables, and relational correlates, that aim to explain the development of eating disorders in women (Leung, Geller, & Katzman, 1996;Mintz & Wright, 1993;Striegel-Moore & Cachelin, 1999, 2001. Stice and Agras (1998) Thompson, 1997;Horn, 1985). Research has consistently found that the sociocultural pressure to be thin and to internalize thin ideals are strong predictors of body image concerns (Ferrand, Champely, & Filaire, 2009;Stice, 1994).
This finding has been supported in research with samples of adolescent females (Pike, 1995) and college aged women (Stice, Ziemba, et al., 1996). Stice, Nemeroff, and Shaw (1996) suggest that these pressures to be thin can be divided further into the internalization of thin ideals and body image disturbances. Women who internalize thin ideals associate being thin with attributes such as happiness, desirability, and socioeconomic status (Tiggemann, 2002). According to Monro and Huon (2006), who used a sample of 72 female university students, the media portrays these idealized images, which result in higher levels of self-objectification that alter eating behaviors.
In addition, Paquette and Raine (2004) found that in a sample of 44 adult women, sociocultural pressures to be thin portrayed by the media were internalized by women's social networks which continued to enforce the societal pressure to be thin and reinforced an unhealthy and unrealistic body norm. Furthermore, Bradford and Petrie (2008), in testing Stice and Agras' (1998) Socioculturally Based Dual-Pathway Model in 236 female college freshman found that levels of internalization of thin ideals were correlated with body dissatisfaction. These findings provide motivation for the current study to determine whether or not sociocultural pressures to be thin and the internalization of these ideals are significant factors when examining body satisfaction among athletes.
Researchers have investigated additional possible correlates and risk factors for body image disturbances. One prospective, longitudinal study (Paxton, Eisenberg, & Neumark-Sztainer, 2006) found that among early adolescent girls (N=440) and boys (N=366), perfectionism was identified as a psychological risk factor for the development of body dissatisfaction. Similarly, Green and Pritchard (2003) developed a model to accurately predict body dissatisfaction in adult men and women, which included perfectionism as an independent variable. Socially prescribed perfectionism has also been found to moderate the effects of body dissatisfaction on bulimic and anorexic symptoms (Brannan & Petrie, 2008). Furthermore, the potentially maladaptive personality trait of perfectionism is prevalent in athletes who often go to extremes to perfect performance (Forsberg & Lock, 2006). Perfectionism may mediate the relationship between the internalization of thin ideals and eating disorder symptomatology. Therefore, the current study will evaluate the empirical contribution of perfectionism to the model being studied.
There is a research divide on whether or not athletes are protected from or are at risk to develop body image problems. More clinical research often concludes that collegiate athletes are at risk for developing body image problems (Anderson, Petrie, & Neumann, 2011;Beals & Manore, 1994;Crissey & Honea, 2006;McClellan, 2009;Smolak, Murnen & Ruble, 2000). These researchers argue that the combination of body image problems and the dietary behaviors of athletes lead to unhealthy eating habits, perhaps due to performance in sports that require low body fat (Goldfield, 2009).
Furthermore, many of the studies that have found that athletes are at greater risk for developing body image concerns suffer from methodological problems. These include an unclear operational definition of body image, measures with poor psychometric properties, and a failure to examine level of competitiveness in athletes (Hausenblas & Downs, 2001). Research conducted with community samples often finds that athletes report more body image satisfaction than control groups (Adame et al., 1991;Hausenblas & Downs, 2001;Mackinnon et al., 2003;Robinson & Ferraro, 2004). This disagreement deserves more research attention that can add to and clarify the debate.
The current study will explore the relationships between these variables as they predict body satisfaction by examining them in groups of athletes and non-athletes of both genders using structural equation modeling analyses to determine the best fitting model for predicting body satisfaction in athletes from independent variables internalization of thin ideals, perfectionism, positive affect, and negative affect.

H1.
It is hypothesized that a lower-order, four factor sociocultural pathway model will best fit the data. To test the theoretical applicability of the sociocultural pathway model in the current sample, this model will contain the variables grounded in previous theory (Stice &Agras, 1998), affect, internalization of thin ideals, perfectionism, and body satisfaction. The four-factor sociocultural pathway model will present good indicators of model fit, including a high CFI (CFI>.90) and low RMSEA (RMSEA<.05).
H2a. It is hypothesized that comparisons of the structural equation model predicting body satisfaction from perfectionism, negative and positive affect and internalization will show support for the predictive model across sex. This model will use maximum likelihood and chi-square differences test to examine model fit and will compare model predictive paths across men and women. It is expected that the model will show good fit to the data and that predictive paths will be similar across men and women.
H2b. As previous research suggests the possibility of sport as protective against body dissatisfaction (Gaspar et al., 2011;Hausenblas & Downs, 2001;Hausenblas & Fallon, 2006;Hausenblas & McNally, 2004), it is hypothesized that comparisons of the structural equation model will not support the model across athlete groups (competitive athletes, recreational athletes, and nonathletes), implying that other theoretical models of body satisfaction should be applied to competitive and recreational athletes.   Johnson, & Diehl, 1996). This scale has two underlying dimensions that measure U.S. societal values concerning attractiveness and beauty.
Nine items evaluated the Importance of Being Physically Fit ("A physically fit and inshape body reflects the beauty ideal for women") and 10 items evaluated the Importance of Being Attractive and Thin ("The thinner a woman is, the more attractive she is"). The BAA-R uses a 7 point Likert scale from 1 (strongly disagree) to 7 (strongly agree). Petrie et al. (1996)   affect, and internalization of thin ideals will significantly predict body satisfaction and will fit these data.  The remaining (n=294) participants were non-athletes. The data from the nonathlete group was used as comparison to the data from the athletes.
Initial data analysis examined means and standard deviations of the overall sample for each demographic variable, original predictor variables, and the original dependent variable (see Table 1). Percentages and standard deviations were examined by sex (Table 2). Chi-squared analyses revealed that race/ethnicity did not vary by sex (X 2 (5)=9.94, p=.08), however athlete group did vary by sex (X 2 (2)=39.31,p<.001). Age was not statistically significantly different between different sex participants (F(26,514)=1.22,p=.22) Percentages and standard deviations were also examined by athlete group (Table 3). Chi-squared analyses revealed that race/ethnicity did vary significantly by athlete group (X 2 (10)=18.58, p<.05). Age was not statistically significantly different across athlete groups (F(26,514)=0.80,p=75). Pearson correlations were also examined for each variable and are presented in Table 4.   Post hoc comparisons using the Tukey HSD test indicated that significant differences were found between competitive and recreational athletes on the perfectionism scale, (M= 2.14, SD=.60), and between competitive and nonathletes, (M=2.83, M=0.57) finding that the more competitive the athlete, the more perfectionistic. Significant differences were also found between recreational and nonathletes on the satisfaction scale, (M=3.64, SD=0.91), finding that the nonathletes showed less satisfaction with their bodies than the recreational athletes. Significant differences were found between recreational and nonathletes on the positive affect subscale (M=-2.53, SD=0.73), finding that recreational athletes had less positive affect than nonathletes. Men had statistically significantly higher scores on perfectionism and body satisfaction than women, and significantly lower scores on negative affect than women, indicating more negative affect. Effect sizes were very small.
Post hoc comparisons using the Tukey HSD test indicated that significant differences were found on the perfectionism scale between competitive and recreational athletes and between competitive and nonathletes, reflecting that more competitive athletes were more perfectionistic. Significant differences were also found on the body satisfaction scale between recreational and nonathletes reflecting that nonathletes showed less satisfaction with their bodies than recreational athletes. Significant differences were also found on the positive affect subscale between recreational and nonathletes reflecting that recreational athletes had less positive affect than nonathletes.

Confirmatory Measurement Models
Confirmatory factor analyses were conducted on the original factor structure for each measure in order to examine the reported factor structures and fit of the sociocultural pathway model constructs in the current sample. Three models, which included two factor-correlated, two-factor uncorrelated, and one factor-models were examined for the Beliefs About Attractiveness Scale, the Negative and Positive Affect Scale and the Body Parts Satisfaction Scale. Model fit was low in each model for all measures, and, as a result, items with factor loadings <.50 were dropped from the measures in order to improve model fit. Using the results from the confirmatory factor analyses, a structural equation model was conducted in order to determine the models' fit to the data. Finally, the predictive structural equation model will be conducted in each group; men vs. women, and competitive vs. recreational vs. nonathlete, in order to examine the models' fit across sex and athlete groups.

Measurement Models
Confirmatory factor analysis (CFA) is often used to test established theories (Meyers et al. 2006). CFA was used to examine previously reported measurement structures' fit in this sample. One factor loading for each factor was arbitrarily set to one in order to standardize the factor loadings. Three models, two-factor correlated, two-factor uncorrelated, and one factor, were compared for measures assessing  Table 7).
Fit indices showed low-adequate model fit for the revised, two-factor correlated model, χ 2 (53)= 289.21, p<.0001, CFI=0.88, RMSEA=.09, SRMR=0.06 (Table 10). In this model, the correlation between the fit and attractive and thin subscales was r=0.92 Two items from the Positive Affect subscale were cut due to low loadings (affect). The original one factor structure of the Negative and Positive Affect Scale was not retained for analyses, as model fit was unacceptably poor. Therefore, a two-factor correlated factor structure was retained for the revised confirmatory factor analysis, and structural equation model. In this model, the correlation between negative and positive affect subscales was r=0.46. Fit indices showed low-adequate model fit, χ 2 (208)= 1147.62, p<.0001, CFI=0.85, RMSEA=.09, SRMR=0.07 (Table 11).
Seven items, three from the overall body parts satisfaction subscale and four from the satisfaction with face subscale were cut from the Body Parts Satisfaction Scale (Satisfact). Fit indices showed good model fit, χ 2 (14)= 230.38, p<.0001, CFI=0.90, RMSEA=.17, SRMR= 0.06. Seven items were cut from the final CFA results (Table   12). Four of these items represented the satisfaction with face subscale. Therefore, in this sample, one factor, the overall body parts satisfaction subscale, formed the factor reflecting Body Satisfaction. To examine hypothesis 1, a structural equation model predicting Satisfaction from Internal, Perfect, Negative Affect, and Positive Affect was examined. Figure 1 illustrates the hypothesized sociocultural pathway model (H1 internalization of thin ideals, negative and positive affect accounted for R 2 =.14, 14% of the variance in body satisfaction.    Initial confirmatory factor analysis results did not support the hypothesis that the sociocultural model constructs would show good fit to the data. After excluding items with poor factor loadings (<.50), the revised scales showed improvements in fit using a revised factor structure, supporting the sociocultural measurement model. Specifically, there were high loadings on each of the measures, with no substantive cross-loadings and good fit for each of the measurement constructs. Two notable findings during the measurement revision process include the a) poor model fit for the original one factor structure of the Negative and Positive Affect Scale, and the b) poor support for a twofactor correlated model for the Body Parts Satisfaction Scale. As a result, the construct originally labeled as "affect" was separated into two distinct constructs: negative and positive affect, supporting a revised two-factor correlated factor structure. Additionally, the latent variable "satisfaction" was measured by one of the two original subscales, the overall body parts satisfaction subscale. Implications for these findings are that scales meant to measure a single, latent variable of affect using positive and negative markers should be evaluated structurally before assuming perfect correlation between positive and negative constructs. Furthermore, results in this sample indicated that body satisfaction was measured by satisfaction with the overall body separate from the face, indicating that a separate measurement construct should be used when evaluating a participant's satisfaction with his/her face.
Thus, it can be concluded that the measurement model using the factors of the sociocultural pathway model were minimally supported and that the results are grounded in the sociocultural pathway theory in which variables such as internalization of thin ideals, perfectionism, and negative and positive affect lead to body image dissatisfaction. According to the SCD model but not evaluated in these data, such a process could lead to the development of a subclinical or clinical eating disorder. The structural equations model evaluated in this study produced factor loadings which revealed that perfectionism accounted for 1.2% of the overall variance in body satisfaction, internalization accounted for 1% of the overall variance in body satisfaction, negative affect accounted for 1.4% of the overall variance in body satisfaction, and that positive affect accounted for 4.8% of the overall variance in body satisfaction. These findings suggest that, consistent with the overall multivariate effect size being small (R 2 =0.14), perfectionism, negative affect, positive affect, and internalization of thin ideals were also small cross-sectional predictors of body satisfaction in male and female athletes and nonathletes, with positive affect showing the strongest predictive path. Positive and negative affect showed the strongest loadings in the model which supports existing theory that affect and disordered eating behaviors/attitudes may have a cyclical relationship. However, longitudinal designs and analyses would be needed to draw this conclusion. Additionally, there is considerably more room to explain variance in these factors than were accounted for by this model. It is possible that other factors such as pressures from coaches and parents, or extrinsic motivations to succeed could be more appropriate variables to include in future versions of this theoretical model. Suggestions for theoretical considerations in expanding the model are discussed later in this chapter.
The second main finding was that the multigroup analysis provided sufficient evidence that the predictive model had adequate fit in each of the groups, and that predictive paths worked similarly across sex and athlete group. This indicates there is evidence for invariance of predictive paths across sex and athlete groups. However, although the model accounted for the same amount of variance (R 2 =0.15) in men and women, there were important differences across the predictive paths. Specifically, differences were noted in the directionality of perfectionism, where the predictive path from perfectionism to satisfaction was negative for men, and for negative affect, where the predictive path from negative affect to satisfaction was negative for women. These differences indicate that for men, less perfectionism was related to more body satisfaction and that, for women, more negative affect was related to lower body satisfaction.
Previous research findings that sport offers a possible protective factor against body dissatisfaction in athletes led to the null hypothesis that the predictive paths would be similar across athlete groups would be rejected; thus, the structural equation model would not be similar across athlete groups. Additionally, previous analyses on the current data sample (Howes, 2009) found that women scored higher on two measures of eating disorder symptoms than men and that female non-athletes scored highest on the measure for bulimia while female competitive athletes scored highest on the measure for anorexia. This difference in mean scores between measures meant to assess two different eating disorders show that the pathways and constructs predicting eating disorders may differ with each disorder and, in this sample, the means differed across athlete groups. These previous results supported the current hypothesis that the predictive paths would not be similar across athlete groups. Results did support this hypothesis. Although initial model fit was poor for competitive athletes, collapsing athlete groups resulted in adequate model fit for athletes, accounting for (R 2 =0.16) of the overall variance in body satisfaction. Model fit was also adequate in nonathletes, however, accounting for less overall variance (R 2 =0.11) in body satisfaction. There were also notable differences in the predictive paths between athletes and nonathletes.
In athletes, perfectionism accounted for 1.4% of the variance in satisfaction as compared to 0.2% for nonathletes; in athletes, negative affect accounted for 2.1% versus nonathletes at 1.6% of the variance in satisfaction; in athletes, positive affect accounted for 2.3% versus nonathletes at 1.8% of the variance in satisfaction; and, in athletes, internalization accounted for 1.6% versus nonathletes at 0.9% of the variance in satisfaction. Each of the predictive paths accounted for more variance in the athlete group when compared to the nonathlete group. Despite the differences in predictive path weights, the directionality of the weights was similar across athletes and nonathletes. Meaning, the more a participant endorsed being perfectionistic, the more they reported being satisfied with their bodies; the less depressed/endorsing negative affect, the more they reported being satisfied with their bodies; the less happy/feelings of positive affect, the more they reported being satisfied with their bodies; and the less internalization of thin ideals, the more they reported being satisfied with their bodies.
Each of these is consistent with the sociocultural theory, except for positive affect which seems to contradict theory that positive affect offers a buffer against body dissatisfaction.
Although the multigroup comparisons of the predictive model supported the similar predictive paths across groups in this sample, the effect size for the structural model in the overall sample was small. The field of body satisfaction and disordered eating behaviors/attitudes in athletes is in need of expanded measurement models to include constructs that more powerfully predict body dissatisfaction. As highlighted in the current research, research found direct relationships between body image dissatisfaction and eating disorder symptoms in athletes. However, Anderson, Petrie, and Neumann (2011) found that body dissatisfaction was unrelated to dietary restraint in their sample of swimmers and gymnasts. This may reflect that subclinical eating disorder symptoms in female athletes participating in weight-conscious sports who may choose to restrict their eating as a result of sport related pressures even if they are satisfied with their bodies Neumann, 2011). Ferrand et al. (2009) found similar results in their sample of rhythmic gymnasts, of which 37.5% were at risk for developing disordered eating, according to EAT-26 scores, yet were satisfied with their appearance. These findings suggest that more research is needed to critically examine models that propose a direct relationship between body image dissatisfaction and disordered eating in athletes. For this reason, the sociocultural model of eating disorder development may be too simplistic in determining how disordered eating behaviors manifest in athletes (Fallon & Hausenblas, 2005). Additionally, such results have implications for future research on disordered eating symptoms after athletes leave the sport, whether due to end of season, injury, or retirement (Papathmoas & Petrie, in press). If an athlete's motivation for weight controlling behaviors is to have a body that will lead to optimal performance and is unrelated to body image dissatisfaction, will those behaviors continue once an athlete is in the off-season, is injured, or retired? Or is it possible that eating disorder behaviors begin during an athlete's career and then become clinically significant over time, especially during retirement, as seen in two studies of retired gymnasts (Kerr, Berman, & De Souza, 2006;Stirling, Cruz, & Kerr, 2012)? Taking into account that involuntary retirement of an athlete can lead to increased anxiety and depression and decreased self-respect and control (Martin, 2014), would athletes entering into retirement be at greater risk for developing clinically relevant eating disorders? Would periods of injury also place athletes at risk given the stress, panic, and fear caused by an injury (Lynch, 1988;Quinn & Fallon, 1999;Tracey, 2003;Williams & Andersen, 1998)? For these reasons, it is strongly suggested that longitudinal research be conducted to determine a) whether or not weight-controlling behaviors carryover into the off season, periods of injury, and retirement and, if so, 2) the role of body image dissatisfaction in subclinical eating disorder symptoms during the off season, periods of injury, and retirement.
Possibly the most important feature of the construct of body satisfaction and disordered eating behaviors/attitudes in athletes is that they have not been operationally defined very well. Many studies argue that athletes might have a higher degree of disordered eating behaviors and attitudes than the nonathlete population, yet these studies use different outcome measures to test the same construct (Garner & Garfinkel, 1980;Hausenblas & Carron, 1999). For example, disordered eating can be evaluated using the non-clinically derived Eating Attitudes Test (Garner & Garfinkel, 1979) and Eating Disorder Inventory (Garner, Olmstead, & Polivy, 1983) or the clinically derived Eating Disorder Examination Interview (Cooper & Fairburn, 1987). Each of these assessments measures symptoms differently. Studies which claim to diagnose or label athletes as "eating disordered" using the EAT or the EDI do so inaccurately since these measures were not constructed using DSM criteria and since clinical diagnoses require clinical interviews. Therefore, without psychometrically sound agreed-upon measures, how can the field draw any conclusions about whether athletes are at greater risk for disordered eating than nonathletes? The lack of one well-developed instrument to measure disordered eating is one explanation for the wide range of prevalence rates of subclinical eating disorder symptoms in athletes ranging from 0%-2% for anorexia nervosa, 1.1%-6% for bulimia nervosa . Thus, the field of sport psychology should develop a psychometrically sound and sensitive measure that can accurately identify body satisfaction and disordered eating symptoms, as they exist specifically in athletes. This measure would be sensitive to the sporting environment, including pressures from coaches and teammates, and would recognize the variability between sports by including sport specific questions (i.e. a thin body in ballet is a sign of professional achievement; wrestlers often need to engage in behaviors to make weight). If such a measure were developed and validated, the field could accumulate both a more accurate understanding of how subclinical eating disorder symptoms manifest in athletes and more precise estimates of prevalence rates.
Similarly, the field is in need of measures that accurately evaluate the construct of body image dissatisfaction (Fiske, Fallon, Blissmer, & Redding, 2014). As demonstrated in the current study, the items that accurately measured the latent variable of body satisfaction needed substantial revision compared to the originally published scale (BPSS-R; Petrie, Tripp, & Harvey, 2002). Sport psychology is also in need of refined and psychometrically sound instruments to accurately measure body dissatisfaction, as suggested by Fiske, Fallon, Blissmer, and Redding (2014) in their review of body dissatisfaction prevalence in U.S. adults.
The current research supports the need to expand body image and eating disorder development models beyond the sociocultural theory to incorporate more expansive models, such as Bronfenbrenner's ecological systems theory (Broderick & Blewitt, 2010). In order to adequately treat an athlete, it is vital that the field recognize the many systems that influence an individual, including families, peers, and the broader society (i.e. the microsystem, mesosystem, exosystem, and macrosystem). It is possible that an ecological approach will also contribute to further understanding of how variables like gender, pressures from society and coaches, age, nationality, and other psychopathology interact and lead to body image dissatisfaction or disordered eating in athletes. Studies by Shanmugam, Jowett, and Meyer (2013) and Karr et al. (2013) have examined both interpersonal and intrapersonal factors in athlete body dissatisfaction and eating psychopathology development and yield promising results for the broadening of models beyond the sociocultural model. Fallon and Hausenblas (2005) previously criticized the sociocultural model for its failure to explain why not all who internalize ideal images develop an eating disorder. Perhaps the incorporation of Bronfenbrenner's ecological model to quantitative and qualitative research on body image dissatisfaction and subclinical eating disorder development in athletes will help tease out the many unanswered questions of their etiology and lead to more refined professional services for the treatment of athletes.

Future Directions and Limitations
Although prevention programs do exist that have empirical support, the field is in need of more research in the area of prevention. Furthermore, prevention efforts have been thwarted by the inability of the field to come to a consensus on one model of disordered eating behaviors/attitudes development. Knowing that researchers conduct and clinicians practice using different theoretical orientations, one model may never emerge as predominant. This is unfortunate as prevention programs are most effective when risk variables can be identified. Nonetheless, adopting a model that includes various systems, such as Bronfenbrenner's Ecological Systems Theory (Broderick & Blewitt, 2010), would allow for a wide variety of previously proven significant variables in explaining the development of disordered eating behaviors/attitudes. The Ecological Systems Theory accounts for the many institutions that influence and impact the psychology of an individual. Perhaps the addition of a genetic or biological component to this theory would provide better causal models that could lend themselves to statistical testing; thus, resulting in refined prevention efforts.
As mentioned previously, the field must consider its gender bias, both in research and in practice, and begin to include, study and treat men with disordered eating behaviors/attitudes as much as it does women. The same is also true for ethnic minorities, and the LGBTQQ population.
In conclusion, it is suggested that prevention programs stay current and incorporate new, scientifically proven, successful interventions into their curriculum.
Specifically, mindfulness-based approaches have been shown to have strong and successful links to the reduction of body dissatisfaction, emotional eating, and dichotomous thinking (Alberts, Thewissen, & Raes 2012). Borrowing from positive psychology, mindfulness emphasizes self-compassion and non-judgment in order to cultivate self-acceptance (Kristeller & Hallet, 1999). These methods could be useful in the prevention of disordered eating behaviors/attitudes as the sociocultural, feminist, and cognitive-behavioral models explaining disordered eating behaviors/attitudes development each contain a component of perfectionism or negative self-evaluation. If a young man or woman can learn how to develop self-compassion and self-acceptance, their self-esteem may rise and they may learn to question external ideals of beauty.
Future research should seek to examine the utility of implementing mindfulness-based techniques in eating disorder prevention programs.
Cross-sectional research on eating disorder development has given the field an acceptable place to start in creating explanatory models. The hope is that, with a better understanding of the predictors or risk factors of disordered eating behaviors/attitudes, prevention programs will be constructed to address each of the variables explaining eating disorder development. The reality, however, is that multiple variables influenced by many different institutions, and their interactions, can lead to disordered eating behaviors/attitudes. Recently, explanatory models have become muddled with the substantial evidence that genetics and biology play a role in whether or not an individual develops an eating disorder (Strober, Peris, & Steiger, 2014). Since our current understanding of how to alter genetics remains very limited, the question must be asked, is it possible to develop a program to actually prevent disordered eating behaviors/attitudes from happening? Therefore, it is proposed that future longitudinal designs be created that integrate theory from positive psychology, specifically the literature around resiliency, to answer the questions of what protective factors can be taught rather than what risk factors can be prevented. Nonetheless, future study designs should focus on identifying variables that protect against disordered eating behaviors/attitudes rather than what might place one individual more at risk.
Limitations of the present study include its inability to test the dual pathway portion of the Socioculturally Based Dual-Pathway Model. Bradford and Petrie (2008) tested the cyclical nature of negative affect and disordered eating attitudes by using a longitudinal design. The current study is cross sectional and as such was not able to draw causal conclusions or conclusions regarding the cyclical relationship of these two constructs. In addition, results were collected using self-report measures, which may be subject to biased responses (Cook & Campbell, 1979). Similarly, due to the secretive nature of disordered eating behaviors/attitudes, it is possible that respondents were not entirely honest with their answers (Bradford & Petrie, 2008). Multiple sample modeling should be conducted in the future to evaluate this model's comparability across subgroups. This study would have benefited from a larger sample of competitive athletes and male participants who identified as either athlete or non-athlete. In addition, it would be useful for future studies to involve groups of male athletes who are at risk for competing in sports where leanness is essential such as wrestlers or jockeys as this data might clarify some of these suggestive findings. You are being asked to participate in a research project that examines different factors that predict disordered eating in individuals. Please read the following information carefully before you decide to complete the questionnaire. If you have any questions, regarding any portion of this research, please feel free to contact the researcher, Sarah Howes, via e-mail at showes@mail.uri.edu. You must be at least 18 years old to participate in this research project Description of the Project: The purpose of this project is to test the socioculturally based dual-pathway model by examining the prediction of disordered eating among athletes and non-athletes. It makes no difference if you are an athlete, a non-athlete, or have ever been diagnosed with an eating disorder.
If you decide to take part in this study, your participation will involve completion of the questionnaire following your approval that you are at least 18 years of age. Your entire time commitment should be about 15-25 minutes.
While every attempt has been made to be sensitive to your experience with body dissatisfaction and possible disordered eating, there is still a possibility that some questions may make you uncomfortable or upset.
Although there may be no direct benefit to you for taking part in this study, your responses provide a tremendous resource in helping to understand the predictors of disordered eating. Research in this area is scarce and your participation is crucial to the advancement of knowledge in this regard.
Your part in this study is confidential and anonymous. None of the information will identify you by name, therefore, no one will know you, personally, participated in this study and no one can find out what your answers were.
Participation in this study is not expected to be injurious or harmful to you; however, if this study causes you an injury, you should write or call Sarah Howes (see above) and/or Charles Collyer, Ph.D. at the University of Rhode Island at (401) 874-4227.
Your participation in this study is strictly voluntary and you can refuse to answer any question. If you decide to take part in this study, you may stop at any time.
Whatever you decide will have no negative consequences for you. If you choose not to participate, you can simply disregard this request.
If you have any questions or concerns about this research, you may discuss your complaints with Sarah Howes (see above) or with Charles Collyer Ph.D. (see above), anonymously, if you choose. In addition, you may contact the University of Rhode Island's office of the Vice President for Research, 70 Lower College Road, Suite 2, University of Rhode Island, Kingston, Rhode Island, telephone: (401) 874-4328.
If for any reason these questions are upsetting to you and you would like to talk to someone, you may contact any of the following organizations for assistance and/or referral: • National Eating Disorders Organization- (206)  If you are at least 18 years old, you have read the consent form and your questions have been answered to your satisfaction, then completion of the survey to follow implies your consent to participate in this study. Please retain this form for yourself. Your signature is not required, since this research is both anonymous and confidential. It is very important that you respond to all these items and that you answer them honestly.