Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: Psychol Sport Exerc. 2020 Jun 18;50:101745. doi: 10.1016/j.psychsport.2020.101745

Development and Validation of the Eating Disorders Screen for Athletes (EDSA): A Brief Screening Tool for Male and Female Athletes

Vivienne M Hazzard a,b, Lauren M Schaefer b, Allison Mankowski c, Traci L Carson a, Sarah M Lipson d,e, Claire Fendrick a, Ross D Crosby b, Kendrin R Sonneville a
PMCID: PMC7392177  NIHMSID: NIHMS1610207  PMID: 32733166

Abstract

Objectives:

To develop and validate the Eating Disorders Screen for Athletes (EDSA), a brief eating disorders screening tool for use in both male and female athletes.

Methods:

Data from Division I athletes at a Midwestern university (N=434) were used to conduct exploratory factor analysis (EFA) by gender. Data from athletes competing at various levels at universities across the United States (N=862) were then used to conduct confirmatory factor analysis (CFA) and receiver operator characteristic (ROC) curve analysis by gender. Athletes from a range of lean and non-lean sports were included. Gender-specific empirically derived cut-offs on the Eating Disorder Examination-Questionnaire were used to classify high eating disorder risk for ROC curve analysis. Measurement invariance by gender, level of competition, and sport type was also examined.

Results:

A six-item, one-factor structure for the EDSA was supported by EFA and CFA in both genders, and internal consistency was good for both male (α=. 80) and female athletes (α=.86). ROC curve analyses indicated that the EDSA was highly accurate in predicting eating disorder risk status and identified a score of 3.33 as the optimal cut-off for both male (sensitivity=.96, specificity=.80) and female athletes (sensitivity=.96, specificity=.64). Results also supported strong measurement invariance for the EDSA by gender, level of competition (Division I versus club), and sport type (lean versus non-lean).

Conclusions:

The EDSA shows promise as a brief screening tool to identify male and female athletes at risk for eating disorders.

Keywords: Athletes, Feeding and Eating Disorders, Psychometrics

Introduction

Eating disorders (e.g., anorexia nervosa, bulimia nervosa, binge-eating disorder) are serious mental disorders that are estimated to affect between 1% - 3% of young adult males and between 6% - 15% of young adult females (Allen et al., 2013; Smink et al., 2014; Stice et al., 2013) according to criteria from the most recent edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association, 2013). Subthreshold disordered eating is more common than full threshold eating disorders, with up to 36% of males and 58% of females reporting disordered eating in young adulthood (Eisenberg et al., 2012), and both full threshold eating disorders and subthreshold disordered eating are associated with substantial psychosocial impairment and medical complications (Ackard et al., 2011; Hudson et al., 2007; Song & Lee, 2019; Udo & Grilo, 2019). Athletes appear to be at increased risk for eating disorders and disordered eating (Bratland-Sanda & Sundgot-Borgen, 2013; Byrne & McLean, 2002; Schaal et al., 2011; Sundgot-Borgen, Jorunn Torstveit, 2004). However, identification of eating pathology in this population is complicated by expectations for athletes to train at high intensities and for long periods of time, as well as to eat in ways that optimize athletic performance (Chapa et al., 2018; Currie, 2010). Importantly, eating pathology is associated with decreased athletic performance (El Ghoch et al., 2013; Mountjoy et al., 2014) and increased risk for sport-related injuries (e.g, stress fractures; Tenforde, Barrack, Nattiv, & Fredericson, 2016), compounding the importance of identifying and intervening upon disordered eating in this population.

In addition to eating disorder risk factors observed among the general population, such as sociocultural influences (e.g., idealized bodies in the media, pressures to attain body ideals, internalization of body ideals) and personality traits (e.g., perfectionism, neuroticism, impulsivity; Culbert, Racine, & Klump, 2015; Lilenfeld, Wonderlich, Riso, Crosby, & Mitchell, 2006; Thompson & Stice, 2001), athletes face a unique set of eating disorder risk factors related to sport participation, such as performance pressure, team weigh-ins, uniforms, and injury (Bratland-Sanda & Sundgot-Borgen, 2013; Greenleaf, 2002). In particular, athletes participating in lean sports – those that emphasize leanness as a means to improve performance (e.g., cross country), are judged aesthetically (e.g., gymnastics), or have weight classes (e.g., wrestling) – tend to have higher eating disorder risk than athletes participating in non-lean sports, such as basketball, golf, and softball (Bratland-Sanda & Sundgot-Borgen, 2013; Joy et al., 2016). Eating disorder risk has also been found to vary by level of competition (Darcy, Hardy, Lock, et al., 2013; Picard, 1999). Additionally, some research suggests that eating disorder risk is higher among female athletes than male athletes (Bratland-Sanda & Sundgot-Borgen, 2013; Joy et al., 2016), similar to the pattern observed in the general population. However, recent studies suggest that the gap in eating disorder risk between male and female athletes may be smaller than previously thought (McDonald et al., 2019; Pustivšek et al., 2019), and both male and female athletes experience a range of adverse health and performance consequences associated with their disordered eating (Joy et al., 2016).

Despite growing recognition of the elevated prevalence and detrimental impact of eating pathology among both male and female athletes, identification of clinically significant disordered eating in this at-risk population poses significant challenges. For example, eating disorder assessment measures designed for use with general populations have demonstrated suboptimal psychometric properties in athlete populations (Pope et al., 2015). This is not surprising, as athletes’ motivations and norms pertaining to eating, exercise, and body image differ from those of the general population (Chapa et al., 2018; Currie, 2010), impacting the interpretation and response patterns for items designed to assess these constructs among non-athletes. Several measures to assess eating disorder risk among female athletes have been developed and validated, including the Athletic Milieu Direct Questionnaire (AMDQ; Nagel et al., 2000), the Female Athlete Screening Tool (FAST; McNulty et al., 2001), the Physiologic Screening Test (PST; Black et al., 2003) and the Brief Eating Disorder in Athletes Questionnaire (BEDA-Q; Martinsen et al., 2014). In addition, although the Compulsive Exercise Test – Athlete Version (CET-A; Plateau et al., 2014; Taranis et al., 2011) was developed to assess excessive exercise, it has been found to successfully discriminate between female athletes with and without high eating disorder risk (Plateau et al., 2017). However, each of these measures was developed and validated for use with female athletes only. Thus, their utility for identifying eating pathology among males, who demonstrate different body ideals and patterns of disordered eating than females (Murray et al., 2017), is limited. Further, with the exception of the BEDA-Q, these measures are relatively lengthy, limiting ease of dissemination as screening tools. Given this, there is a critical need in the field for a brief psychometrically sound screening tool that could be used to identify male and female athletes with elevated eating disorder risk. Therefore, the primary objective of the present study was to develop and validate the Eating Disorders Screen for Athletes (EDSA), a brief eating disorders screening tool for use in both male and female athletes. In addition, the present study aimed to identify clinical cut-offs for males and females and to evaluate the appropriateness of comparing EDSA scores across distinct groups of athletes via measurement invariance testing by gender, level of competition, and sport type.

General Method

Two studies were conducted to develop and validate the EDSA. Study 1 used data from a sample of Division I athletes at a single Midwestern university to examine the factor structure of the EDSA. Study 2 used data from athletes competing at various levels at universities across the United States to examine the factorial validity, internal consistency, criterion validity, and measurement invariance of the EDSA.

Item Generation

The EDSA development resulted from a research-practice collaboration that initially began with twenty disordered eating items from the FAST (McNulty et al., 2001) and the Female Athlete Triad Screening Questionnaire (Mountjoy et al., 2015). While both tools were designed for use with female athletes, the twenty items selected were chosen for potential relevance among male athletes as well. However, a study by Gallagher and colleagues (2019) identified considerable gender bias in several of these items, such that female athletes were more likely than male athletes to endorse items focused on thinness and weight even after adjusting for overall eating disorder risk. Therefore, we modified items to focus on more gender-neutral body ideals (e.g., leanness rather than thinness; Smolak & Murnen, 2008) in order to reduce gender bias. For example, we modified the FAST item, “Do you worry that you will gain weight if you cannot exercise?” (McNulty, Adams, Anderson, & Affenito, 2001) to, “Do you worry that your weight, shape, or body composition will change if you cannot exercise?”.

Six of these modified items (see Appendix) administered in both Study 1 and Study 2 comprise a short but comprehensive list of items representing adequate construct coverage of the core attitudinal features of eating disorders (Cooper & Fairburn, 1987). Two of these items originated from the FAST (McNulty et al., 2001) and represent weight/shape concerns; four of these items originated from disordered eating items on the Female Athlete Triad Screening Questionnaire (Mountjoy et al., 2015) and represent importance of weight, weight concerns, binge eating-related concerns, and dietary restraint. Additional items beyond these six were piloted but not administered in both samples. Therefore, in order to conduct exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) in separate, adequately sized samples, as is considered best practice (Swami & Barron, 2019; Worthington & Whittaker, 2006), only these six items were included in analyses for the studies described below.

Study 1: Scale Development and Identification of Scale Structure

Method

Participants and Procedure

Male-identifying (N = 229) and female-identifying (N = 205) Division I athletes at a Midwestern university completed the six-item EDSA online as part of their annual fall pre-participation screening. Due to the nature of assessment, responses were not anonymous, and all responses were required; thus, there were no missing data. Athletes represented 11 different sports, including lean (i.e., cross country, gymnastics, rowing, swim and dive, and track and field) and non-lean (i.e., baseball, basketball, football, golf, soccer, and volleyball) sports. The study was determined exempt by the governing Institutional Review Board because this research used de-identified existing data after it was collected for the purpose of fall pre-participation screening, and participant consent was waived by the approving Institutional Review Board.

Measures

Athletes completed the six EDSA items. Responses were recorded on a 5-point Likert-type frequency scale with response options of 1 = never, 2 = rarely, 3 = sometimes, 4 = often, and 5 = always.

Statistical Analysis

Parallel analysis and EFA were conducted separately among male and female athletes using SPSS 25 to identify the initial factor structure of the EDSA in each group. In EFA, Bartlett’s test of sphericity and the Kaiser-Meyer-Olkin measure of sampling adequacy were used to assess the factorability of the EDSA items. Items are considered appropriate for factor analysis when Bartlett’s test is statistically significant and the Kaiser-Meyer-Olkin value is ≥ 0.60 (Tabachnick & Fidell, 2006). The number of factors to retain was determined via examination of the scree plot and parallel analysis (Horn, 1965), in which eigenvalues from the observed data were compared with 95th percentile eigenvalues from 1,000 permutations of the original dataset (O’Connor, 2000). EFA was conducted using principal axis factoring and promax oblique rotation. Primary loadings ≤ .40 (Johnson & Morgan, 2016; Stevens, 1992) and cross-loadings > .40 (Hooper, 2012) in EFA, as well as corrected item-to-total correlations < .40 (Lienert & Raatz, 1998; Nunnally & Bernstein, 1994), were used as criteria for item deletion.

Results

Exploratory Factor Analysis

Bartlett’s test of sphericity was significant for males (χ2 = 381.79, df = 15, p < .001) and females (χ2 = 601.48, df = 15, p < .001), and the Kaiser-Meyer-Olkin value was .82 for males and .87 for females, indicating the EDSA items were appropriate for factor analysis in each gender. As shown in Figure 1, examination of the scree plot and parallel analysis indicated a one-factor solution for both male and female athletes. EFA results by gender are presented in Table 1. All primary loadings in EFA were > .40, and corrected item-to-total correlations ranged from .40-.66 among male athletes and from .60-.75 among female athletes; thus, no items were deleted.

Figure 1.

Figure 1.

Open in a new tab

Scree and parallel analysis plots for male and female athletes. The solid red lines represent eigenvalues from the observed data; the dashed blue lines represent 95th percentile eigenvalues from 1,000 permutations of the original data.

Table 1.

Summary of Exploratory Factor Analysis Results

Male Athletes Female Athletes
Factor 1
Eigenvaluea 2.99 3.77
Percent variance explained 49.89 62.85
Item Factor Loading
Does your weight, shape, or body composition affect the way you feel about yourself? .76 .83
Are you dissatisfied with your weight, shape, or body composition? .64 .81
Do you worry that your weight, shape, or body composition will change if you cannot exercise? .78 .75
Do you want to be leaner even if others may think you are already lean? .43 .74
Do you worry about losing control over your eating because of how it may affect your weight, shape, or body composition? .60 .70
Do you try to avoid certain foods to influence your weight, shape, or body composition? .56 .63
Open in a new tab

Study 2: Confirmation of Factor Structure and Examination of Reliability, Criterion Validity, and Measurement Invariance

Method

Participants and Procedure

Athletes competing at 48 universities across the United States at either the Division I, Division II, Division III, or club level were invited to participate. Universities were selected to represent all major geographic regions across the contiguous United States (i.e., Northeast, Southeast, Midwest, Southwest, and West). Depending on the university and level of competition, athletes were identified either via athletic department administration, club sports leadership, or team rosters, and athletes were invited to participate via email sent by athletic department administration, club sports leadership, or the researchers. Invitation emails described the study as research that could help improve the health and performance of athletes, explained that survey responses would be anonymous, and included an anonymous survey link. Athletes who elected to participate could enter a drawing to win one of ten $100 gift cards as an incentive for participation. Of the 962 respondents that provided informed consent, 21 did not complete any survey items. Of the remaining 941 respondents, we excluded those who did not confirm being an athlete (n = 48), did not report responding truthfully to all survey questions (n = 8), did not identify as male or female (n = 32), or did not complete all EDSA items (n = 22).

The final analytic sample was comprised of 312 male-identifying and 550 female-identifying athletes. The majority of respondents reported competing at the Division I (n = 588) or club (n = 257) level, and there was a relatively equal representation of lean (n = 418) and non-lean (n = 444) sports. The majority (74.1%) of the sample identified as non-Hispanic white. The mean age of athletes was 20.16 years (SD = 1.47), and the mean body mass index (BMI) of athletes calculated using self-reported height and weight was 23.71 kg/m2 (SD = 3.52). The governing Institutional Review Board determined this study exempt because it collected de-identified survey data.

Measures

In addition to the six EDSA items, athletes completed the 22 attitudinal items of the Eating Disorder Examination-Questionnaire (EDE-Q; Fairburn & Beglin, 1994), which assesses eating pathology over the past four weeks. Likert-type items are rated on a 7-point scale ranging from 0 (no days/not at all) to 6 (everyday/markedly). These items are averaged to provide four subscale scores: Restraint, Eating Concern, Shape Concern, and Weight Concern. The global score is calculated as the average of the four subscales. The global score was used to represent eating disorder risk in the present study, with higher scores indicating greater eating disorder risk. In samples of U.S. college students, the EDE-Q global scale has demonstrated excellent internal consistency among males (α = .93) and females (α = .95; Quick & Byrd-Bredbenner, 2013), as well as strong test-retest reliability for males (rho = .89) and females (rho = .90; Rose et al., 2013). The EDE-Q global scale also demonstrated excellent internal consistency in the present sample among male (α = .94) and female (α = .96) athletes. Compensatory behaviors were also assessed by asking athletes whether or not they had vomited or used laxatives in the past four weeks to influence their weight, shape, or body composition.

Statistical Analysis

CFA by gender and measurement invariance testing were conducted in Mplus 8. Only groups that met the suggested minimum sample size of 200 (Kelloway, 2015) were included in multi-group CFAs. Thus, multi-group CFAs utilized the full sample for evaluation of measurement invariance by gender (male versus female) and sport type (lean versus non-lean), but only Division I and club athletes were included for evaluation of measurement invariance by level of competition. Estimation was conducted via weighted least squares with mean and variance adjustment (WLSMV), as has been recommended for ordinal data (Brown, 2015). SPSS 25 was used to compute internal consistency and descriptive statistics, as well as to conduct receiver operating characteristic (ROC) curve analysis by gender to assess criterion validity and identify a clinical cut-off for the EDSA.

Confirmatory Factor Analysis.

The EDSA model identified via EFA was evaluated with CFA. Adequacy of model fit was judged by the following: comparative fit index (CFI), root mean square error of approximation (RMSEA), and standardized root-mean square residual (SRMR). Values ≥ .95 for CFI, ≤ .06 for RMSEA, and < .08 for SRMR indicate good model fit (Hu & Bentler, 1999). Values of .90 or higher for CFI, up to .10 for RMSEA, and up to .10 for SRMR indicate acceptable but mediocre model fit (Bentler, 1990; Browne & Cudeck, 1993; Hu & Bentler, 1995; MacCallum, Browne, & Sugawara, 1996; Schermelleh-Engel & Müller, 2003). Models were deemed to have adequate fit if most fit indices suggested acceptable fit.

Receiver Operating Characteristic Curve Analysis.

Gender-specific empirically derived cut-offs for the EDE-Q global score of 1.7 for males (Schaefer et al., 2018) and 2.3 for females (Mond et al., 2004) were used to classify eating disorder risk for ROC curve analysis, resulting in 53 (17.0%) males and 155 (28.2%) females classified as high risk. ROC curve analysis was used to assess the accuracy of the EDSA score (calculated by averaging responses to the items in the final model) in predicting eating disorder risk status, as well as to identify a cut-off at which the EDSA score provided the optimal trade-off between sensitivity and specificity. Given the goal of using the EDSA as a screening tool, sensitivity was given stronger consideration than specificity.

External Validation Using Compensatory Behavior Data.

Using chi-square tests, compensatory behavior use was compared across eating disorder risk status groups defined by the EDSA cut-off identified in ROC curve analysis.

Measurement Invariance Testing.

The EDSA model underwent multi-group CFAs to assess configural, metric, and scalar measurement invariance across groups by gender (male versus female), level of competition (Division I versus club), and sport type (lean versus non-lean). The first step in each multi-group CFA assessed configural invariance, in which all factor loadings and item thresholds were free to vary across groups. The second step assessed metric (i.e., weak) invariance, in which the factor loadings were constrained to be equal across groups. The third step assessed scalar (i.e., strong) invariance, in which factor loadings and item thresholds were constrained across groups. Scalar invariance indicates that levels of the latent factors indexed by a given scale/subscale have the same meaning across groups, and therefore group factor scores can be meaningfully compared (Brown, 2015). Nested models (i.e., metric compared to configural; scalar compared to metric) were compared using the changes in CFI, RMSEA, and SRMR, as well as the DIFFTEST function in Mplus, which is the χ2 difference test for WLSMV estimation. Because χ2 is sensitive to sample size, changes in CFI, RMSEA, and SRMR were given stronger consideration than the χ2 difference test (Putnick & Bornstein, 2016). Specifically, significant decrement in fit across nested models was indicated by changes greater than .010 for CFI, .015 for RMSEA, and .030 for SRMR for metric invariance or .010 for SRMR for scalar invariance (Chen, 2007; Cheung & Rensvold, 2002). As the stepwise progression from configural to scalar invariance places increasing equality constraints across a family of parameters in the CFA model (e.g., all item thresholds assumed to be equal), significant decrements in model fit between these nested models indicates that at least one parameter is noninvariant (i.e., full invariance is not supported) but should not be interpreted to mean that all parameters are noninvariant. Fit diagnostics (e.g., modification indices) can be used to identify the source of misfit and test for partial measurement invariance. Partial invariance is supported when some but not all parameters are equivalent. Research suggests that scale means can be meaningfully compared across groups if partial metric and partial scalar invariance are observed (Brown, 2015; Byrne, Shavelson, & Muthén, 1989).

Results

Confirmatory Factor Analysis

CFA indicated adequate model fit for the six-item, one-factor EDSA model identified via EFA in male athletes (CFI = .972, RMSEA = .104 with 90% CI = .073-.138, SRMR = .034) and female athletes (CFI = .986, RMSEA = .096 with 90% CI = .073-.121, SRMR = .026).

Internal Consistency

Internal consistency of the EDSA was good for both male athletes (α = .80) and female athletes (α = .86).

Criterion Validity

As shown in Figure 2, EDSA scores demonstrated excellent accuracy in predicting eating disorder risk status as determined via EDE-Q for both male athletes (area under the curve [AUC] = .94, 95% CI: .91-.98) and female athletes (AUC = .92, 95% CI: .89-.94). A cut-off score of 3.33 on the EDSA provided optimal trade-off between sensitivity and specificity across genders, resulting in a sensitivity of .96 and specificity of .80 for male athletes and a sensitivity of .96 and specificity of .64 for female athletes.

Figure 2.

Figure 2.

Open in a new tab

Receiver operating characteristic curves for male and female athletes. Area under the curve = .94 for male athletes and .92 for female athletes.

Compensatory Behaviors by Risk Status

Differences in compensatory behavior use by eating disorder risk status as determined by the EDSA cut-off of 3.33 are presented in Table 2. In both male and female athletes, athletes classified as at-risk by the EDSA were more likely than athletes classified as not-at-risk to report vomiting and laxative use.

Table 2.

Compensatory Behaviors by Eating Disorder Risk Status as Determined by the EDSA

Male Athletes
 Female Athletes
Not at Risk
(EDSA < 3.33)
N = 213		 At Risk
(EDSA ≥ 3.33)
N = 99		 Not at Risk
(EDSA < 3.33)
N = 259		 At Risk
(EDSA ≥ 3.33)
N = 291
% (n) p % (n) p
Vomiting 0.5 (1) 8.1 (8) < .001 0.0 (0) 9.8 (28) < .001
Laxatives 0.0 (0) 2.0 (2) .04 0.0 (0) 4.9 (14) < .001
Open in a new tab

Note. EDSA = Eating Disorders Screen for Athletes. Compensatory behaviors used to influence weight, shape, or body composition were assessed over the past four-week period.

Measurement Invariance

Measurement invariance results by gender, level of competition, and sport type from multi-group CFAs are reported in Table 3. Several ΔRMSEA values were above the .015 threshold. In all of these instances, however, the ΔRMSEA corresponded to improved model fit with the addition of parameter constraints, as is expected given that the RMSEA statistic rewards model parsimony (Brown, 2015), which is achieved at increasing levels of invariance testing. Configural, metric, and scalar measurement invariance were supported across groups by level of competition and sport type. Full scalar invariance of the EDSA across genders was not supported, as evidenced by a ΔCFI of .014 indicating worse model fit with the addition of item threshold constraints. However, modification indices suggested relaxing the threshold constraint for the item “Do you try to avoid certain foods to influence your weight, shape, or body composition?”, and when this item threshold was allowed to vary, results supported partial scalar invariance (ΔCFI = .003). For this item, the thresholds were larger for female (−1.04, −0.21, 0.63, 1.60) than male (−1.18, −0.53, 0.17, 1.15) athletes. Therefore, for athletes with the same overall EDSA score, females tended to report avoiding certain foods slightly more frequently than males.

Table 3.

Measurement Invariance by Gender, Level of Competition, and Sport Type

Invariance Model χ2 (df) CFI RMSEA (90% CI) SRMR Δχ2 (Δdf) ΔCFI ΔRMSEA ΔSRMR Invariance?
Testing invariance by gender: male (n = 312) vs. female (n = 550)
Configural 93.76 (18)*** .983 .099 (.080, .119) .029 -- -- -- -- Yes
Metric 88.85 (23)*** .986 .082 (.064, .100) .030 8.22 (5) .003 .017 .001 Yes
Scalar 166.87 (40)*** .972 .086 (.073, .099) .034 82.56 (17)*** .014 .004 .004 No
Testing invariance by level of competition: Division I (n = 588) vs. club (n = 257)
Configural 90.73 (18)*** .984 .098 (.078, .118) .028 -- -- -- -- Yes
Metric 73.92 (23)*** .989 .072 (.054, .091) .029 1.59 (5) .005 .026 .001 Yes
Scalar 86.75 (40)*** .990 .053 (.037, .068) .030 17.46 (17) .001 .019 .001 Yes
Testing invariance by sport type: lean (n = 418) vs. non-lean (n = 444)
Configural 97.32 (18)*** .984 .101 (.082, .121) .028 -- -- -- -- Yes
Metric 86.79 (23)*** .987 .080 (.063, .099) .029 5.94 (5) .003 .021 .001 Yes
Scalar 119.11 (40)*** .984 .068 (.054, .082) .033 38.40 (17)** .003 .012 .004 Yes
Open in a new tab

Note. CFI = comparative fit index; RMSEA = root mean square error of approximation; CI = confidence interval; SRMR = standardized root-mean square residual.

*

p < .05

**

p < .01

***

p < .001

Mean Scores

The mean EDSA score in the full sample was 3.14 (SD = 0.82). Female athletes (M = 3.29, SD = 0.79) had higher EDSA scores than male athletes (M = 2.86, SD = 0.81), t(860) = 7.56, p < .001, and club athletes (M = 3.24, SD = 0.78) had higher EDSA scores than Division I athletes (M = 3.08, SD = 0.83), t(843) = 2.76, p =.006. EDSA scores did not differ across lean and non-lean sports, t(860) = 1.19, p = .24.

Discussion

Although eating disorders are associated with decreased athletic performance (El Ghoch et al., 2013) and a host of other medical and psychiatric consequences (Ackard et al., 2011; Hudson et al., 2007; Mitchell & Crow, 2006), the majority of athletic departments in the United States do not screen for eating disorders (Kroshus, 2016), possibly because current screening tools are limited by their length and focus on female athletes. Therefore, the primary objective of the current study was to develop and validate the EDSA, a brief eating disorders screening tool designed for use with both male and female athletes. Results of exploratory and confirmatory factor analyses in separate samples supported a six-item, one-factor structure of the EDSA among both male and female athletes. Further, the EDSA demonstrated good internal consistency and excellent accuracy in predicting eating disorder risk status as determined via the EDE-Q among both male and female athletes. Examination of sensitivity and specificity values suggest that an EDSA cut-off score of 3.33 best differentiates between athletes with and without high eating disorder risk. Risk status defined by this cut-off was found to be a strong predictor of vomiting and laxative use among both male and female athletes, providing further evidence for validity of the EDSA.

In addition to developing and validating the EDSA, the present study aimed to evaluate the appropriateness of comparing EDSA scores across distinct groups of athletes via measurement invariance testing by gender, level of competition, and sport type. Results supported full strong measurement invariance of the EDSA across level of competition (Division I versus club) and sport type (lean versus non-lean), as well as partial strong measurement invariance across gender (male versus female), indicating that EDSA scores can be meaningfully compared across these groups. As such, EDSA means were compared in the present study. Differences in eating disorder risk as measured by the EDSA existed by gender and level of competition in the present study, such that female athletes appeared to be at higher risk than male athletes, and club athletes appeared to be at higher risk than Division I athletes. The differences observed by gender are consistent with previous literature (Bratland-Sanda & Sundgot-Borgen, 2013; Joy et al., 2016), but prior literature is less clear regarding differences in eating disorder risk across levels of competition. Some studies have found lower levels of competition to be associated with higher eating disorder risk as observed in the current study (Darcy, Hardy, Lock, et al., 2013; Hopkinson & Lock, 2004), while other research has found higher levels of competition to be associated with higher eating disorder risk (Picard, 1999), and yet other research has found no evidence of differences across levels of competition (Holm-Denoma et al., 2009). However, the present sample is not – and was not intended to be – a representative sample of athletes. Therefore, differences observed between club and Division I athletes in the present study may be an artifact of selection bias (e.g., motivations to participate in the survey may have differed between Division I and club athletes) and should be interpreted with caution.

A notable strength of the present study was that, consistent with recommendations for psychometric investigations (Swami & Barron, 2019; Worthington & Whittaker, 2006), exploratory and confirmatory factor analyses were conducted in separate samples to rigorously evaluate the factor structure of the EDSA and identify a model that replicates across samples. Another strength of this study was the relatively large size of each sample, which provided the opportunity to conduct analyses separately by gender. However, the present study had important limitations as well.

An important limitation of this study was the limited pool of items used to develop the EDSA. Additional items beyond the six included here had been considered but were not administered in both samples; therefore, they were not available for both exploratory and confirmatory factor analyses, and neither of our samples were large enough to split into crossvalidation samples. A limitation of Study 1 was potential underreporting of eating pathology, as athletes completed the EDSA as part of their annual fall pre-participation screening and responses were not anonymous. While it is possible that participants in Study 1 may have underreported their disordered eating in order to avoid perceived repercussions (e.g., suspension from sport participation), it is important to note that the EDSA was specifically designed to be used in similar circumstances in order to aid the identification of at-risk athletes by training personnel. Therefore, the finding that the one-factor EDSA structure was replicated across non-anonymous (Study 1) and anonymous (Study 2) reporting conditions is encouraging and demonstrates the robustness of the factor structure. Nonetheless, future work may seek to examine whether a different – perhaps lower – EDSA cut-off may be more appropriate in non-anonymous contexts where underreporting may be a stronger concern. A key limitation of Study 2 was the use of the EDE-Q to assess criterion validity. The EDE-Q provides valuable information about which athletes may be at high risk for an eating disorder; however, unlike a clinical interview, it is unable to provide an eating disorder diagnosis. Further, prior research examining the factor structure of the EDE-Q among athletes suggests the EDE-Q may perform better in athlete populations if some items are dropped (e.g., desire for a flat stomach among males, fear of weight gain among females; Darcy, Hardy, Crosby, et al., 2013). However, in the present study, all EDE-Q items were included because clinically relevant cut-offs for the abbreviated athlete-specific versions of the EDE-Q have not been identified. It is worth noting that although Mond and colleagues (2004) found that using two behavioral items in conjunction with the EDE-Q global score cut-off of 2.3 increased specificity (i.e., correctly ruling out noncases) beyond using the global score cut-off alone, they also found that incorporating the two behavioral items resulted in a decrease in sensitivity from .92 to .83. Therefore, in line with the goal of using the EDSA for screening purposes, we elected to use only the global score cut-off to define eating disorder risk status in the present study, prioritizing sensitivity over specificity. However, this may mean that the already low specificity corresponding to an EDSA cut-off of 3.33 among women (.64) may be an underestimate. In addition, the EDSA cut-off of 3.33 identified in Study 2 was not validated in an independent sample, and although the EDSA is expected to offer improved detection of eating pathology among male and female athletes compared with existing measures which have largely been developed for use with women, the current study was not able to assess the incremental validity of the EDSA. Another limitation of this study was the cross-sectional nature of the data, which did not allow for examination of test-retest reliability. Future research to address these limitations would be valuable to better understand the performance of the EDSA in screening for eating disorder risk. In particular, we recommend evaluating the ability of the EDSA and other measures of eating pathology developed specifically for athletes (e.g., the FAST; McNulty et al., 2001) to predict eating disorder diagnostic status as assessed by clinical interview (e.g., the Eating Disorder Examination Interview; Cooper & Fairburn, 1987).

Conclusions

The EDSA shows promise as a brief screening tool to identify male and female athletes at risk for eating disorders. Implementing the EDSA to screen for and manage eating disorders among athletes has the potential to help improve high risk athletes’ mental health, physical health, and athletic performance.

Highlights.

  • A brief eating disorders screening tool was developed for male and female athletes

  • The Eating Disorders Screen for Athletes (EDSA) is comprised of 6 items

  • Internal consistency was good for male and female athletes

  • The EDSA was highly accurate in predicting eating disorder risk status

  • The EDSA was invariant across gender, level of competition, and sport type

Acknowledgments

Funding: This research was supported by the University of Michigan Library and the National Institute of Mental Health (grant number T32 MH082761).

Abbreviations:

FAST

Female Athlete Screening Tool

BEDA-Q

Brief Eating Disorder in Athletes Questionnaire

EDSA

Eating Disorders Screen for Athletes

EFA

exploratory factor analysis

BMI

body mass index

EDE-Q

Eating Disorder Examination-Questionnaire

CFA

confirmatory factor analysis

WLSMV

weighted least squares with mean and variance adjustment

ROC

receiver operating characteristic

CFI

comparative fit index

RMSEA

root mean square error of approximation

SRMR

standardized root-mean square residual

Appendix

Eating Disorders Screen for Athletes (EDSA)

Please read each question carefully and select the appropriate response. Please note that “weight” refers to numbers on a scale, “shape” refers to amount and distribution of body fat and muscle, “body composition” refers to ratio of body fat to muscle, and “leanness” refers to low body fat-to-muscle ratio.

Never Rarely Sometimes Often Always
1. Does your weight, shape, or body composition affect the way you feel about yourself? 1 2 3 4 5
2. Are you dissatisfied with your weight, shape, or body composition? 1 2 3 4 5
3. Do you worry that your weight, shape, or body composition will change if you cannot exercise? 1 2 3 4 5
4. Do you want to be leaner even if others may think you are already lean? 1 2 3 4 5
5. Do you worry about losing control over your eating because of how it may affect your weight, shape, or body composition? 1 2 3 4 5
6. Do you try to avoid certain foods to influence your weight, shape, or body composition? 1 2 3 4 5
Open in a new tab

Scoring: Average responses across the 6 items. A score ≥ 3.33 indicates high risk for an eating disorder.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interest statement: The authors have no conflicts of interest or financial conflicts relevant to this article to disclose.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

References

  1. Ackard DM, Fulkerson JA, & Neumark-Sztainer D (2011). Psychological and behavioral risk profiles as they relate to eating disorder diagnoses and symptomatology among a school-based sample of youth. International Journal of Eating Disorders, 44(5), 440–446. 10.1002/eat.20846 [DOI] [PubMed] [Google Scholar]
  2. Allen KL, Byrne SM, Oddy WH, & Crosby RD (2013). DSM-IV-TR and DSM-5 eating disorders in adolescents: Prevalence, stability, and psychosocial correlates in a population-based sample of male and female adolescents. Journal of Abnormal Psychology, 122(3), 720–732. 10.1037/a0034004 [DOI] [PubMed] [Google Scholar]
  3. American Psychiatric Association. (2013). Feeding and Eating Disorders In Diagnostic and Statistical Manual of Mental Disorders (DSM-5®). American Psychiatric Association Publishing; [Google Scholar]
  4. Bentler PM (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107(2), 238–246. [DOI] [PubMed] [Google Scholar]
  5. Black DR, Larkin LJS, Coster DC, Leverenz LJ, & Abood DA (2003). Physiologic Screening Test for eating disorders/disordered eating among female collegiate athletes. Journal of Athletic Training, 38(4), 286–297. [PMC free article] [PubMed] [Google Scholar]
  6. Bratland-Sanda S, & Sundgot-Borgen J (2013). Eating disorders in athletes: Overview of prevalence, risk factors and recommendations for prevention and treatment. European Journal of Sport Science, 13(5), 499–508. 10.1080/17461391.2012.740504 [DOI] [PubMed] [Google Scholar]
  7. Brown TA (2015). Confirmatory Factor Analysis for Applied Research, Second Edition Guilford Press. [Google Scholar]
  8. Browne MW, & Cudeck R (1993). Alternative ways of assessing model fit In Bollen KA Long JS (Eds.), Testing Structural Equation Models. Sage. [Google Scholar]
  9. Byrne BM, Shavelson RJ, & Muthén B (1989). Testing for the equivalence of factor covariance and mean structures: The issue of partial measurement invariance. Psychological Bulletin, 105(3), 456–466. 10.1037/0033-2909.105.3.456 [DOI] [Google Scholar]
  10. Byrne S, & McLean N (2002). Elite athletes: Effects of the pressure to be thin. Journal of Science and Medicine in Sport, 5(2), 80–94. 10.1016/S1440-2440(02)80029-9 [DOI] [PubMed] [Google Scholar]
  11. Chapa DAN, Hagan KE, Forbush KT, Perko VL, Sorokina DA, Alasmar AY, Becker CB, Sherman RT, Thompson RA, Farrell JG, & Stewart TM (2018). The Athletes’ Relationships with Training scale (ART): A self-report measure of unhealthy training behaviors associated with eating disorders. International Journal of Eating Disorders, 51(9), 1080–1089. 10.1002/eat.22960 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Chen FF (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling, 14(3), 464–504. 10.1080/10705510701301834 [DOI] [Google Scholar]
  13. Cheung GW, & Rensvold RB (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9(2), 233–255. 10.1207/S15328007SEM0902_5 [DOI] [Google Scholar]
  14. Cooper Z, & Fairburn C (1987). The eating disorder examination: A semi-structured interview for the assessment of the specific psychopathology of eating disorders. International Journal of Eating Disorders, 6(1), 1–8. [DOI] [Google Scholar]
  15. Culbert KM, Racine SE, & Klump KL (2015). Research Review: What we have learned about the causes of eating disorders - a synthesis of sociocultural, psychological, and biological research. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 56(11), 1141–1164. https://doi.org/10.im/jcpp.12441 [DOI] [PubMed] [Google Scholar]
  16. Currie A (2010). Sport and eating disorders - understanding and managing the risks. Asian Journal of Sports Medicine, 1(2), 63–68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Darcy AM, Hardy KK, Crosby RD, Lock J, & Peebles R (2013). Factor structure of the Eating Disorder Examination Questionnaire (EDE-Q) in male and female college athletes. Body Image, 10(3), 399–405. 10.1016/j.bodyim.2013.01.008 [DOI] [PubMed] [Google Scholar]
  18. Darcy AM, Hardy KK, Lock J, Hill KB, & Peebles R (2013). The Eating Disorder Examination Questionnaire (EDE-Q) among university men and women at different levels of athleticism. Eating Behaviors, 14(3), 378–381. 10.1016/j.eatbeh.2013.04.002 [DOI] [PubMed] [Google Scholar]
  19. Eisenberg ME, Berge JM, Fulkerson JA, & Neumark-Sztainer D (2012). Associations between hurtful weight-related comments by family and significant other and the development of disordered eating behaviors in young adults. Journal of Behavioral Medicine, 35(5), 500–508. 10.1007/s10865-011-9378-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. El Ghoch M, Soave F, Calugi S, & Dalle Grave R (2013). Eating disorders, physical fitness and sport performance: A systematic review. Nutrients, 5(12), 5140–5160. 10.3390/nu5125140 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Fairburn C, & Beglin S (1994). Assessment of eating disorders: interview or self-report questionnaire? International Journal of Eating Disorders, 16(4), 363–370. [PubMed] [Google Scholar]
  22. Gallagher KA, Sonneville KR, Hazzard VM, Carson TL, & Needham BL (2019). Evaluating gender bias in an eating disorder risk assessment questionnaire for athletes. Eating Disorders, 1–13. 10.1080/10640266.2019.1613846 [DOI] [PubMed] [Google Scholar]
  23. Greenleaf C (2002). Athletic body image: Exploratory interviews with former competitive female athlete. Women in Sport and Physical Activity Journal, 11(1), 63–88. 10.1123/wspaj.11.1.63 [DOI] [Google Scholar]
  24. Holm-Denoma JM, Scaringi V, Gordon KH, Van Orden KA, & Joiner TE (2009). Eating disorder symptoms among undergraduate varsity athletes, club athletes, independent exercisers, and nonexercisers. International Journal of Eating Disorders, 42(1), 47–53. 10.1002/eat.20560 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Hooper D (2012). Exploratory Factor Analysis In Chen H (Ed.), Approaches to Quantitative Research – Theory and its Practical Application: A Guide to Dissertation Students. Oak Tree Press. [Google Scholar]
  26. Hopkinson RA, & Lock J (2004). Athletics, perfectionism, and disordered eating. Eating and Weight Disorders, 9(2), 99–106. [DOI] [PubMed] [Google Scholar]
  27. Horn JL (1965). A rationale and test for the number of factors in factor analysis. Psychometrika, 30, 179–185. 10.1007/BF02289447 [DOI] [PubMed] [Google Scholar]
  28. Hu L, & Bentler P (1995). Evaluating model fit In Hoyle RH (Ed.), Structural Equation Modeling. Concepts, Issues, and Applications. Sage. [Google Scholar]
  29. Hu LT, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  30. Hudson JI, Hiripi E, Pope HG, & Kessler RC (2007). The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 61(3), 348–358. 10.1016/j.biopsych.2006.03.040 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Johnson RL, & Morgan GB (2016). Survey Scales: A Guide to Development, Analysis, and Reporting. Guilford Publications. [Google Scholar]
  32. Joy E, Kussman A, & Nattiv A (2016). 2016 update on eating disorders in athletes: A comprehensive narrative review with a focus on clinical assessment and management. British Journal of Sports Medicine, 50(3), 154–162. 10.1136/bjsports-2015-095735 [DOI] [PubMed] [Google Scholar]
  33. Kelloway EK (2015). Using Mplus for Structural Equation Modeling: A Researcher’s Guide (2nd ed). SAGE Publications, Inc. [Google Scholar]
  34. Kroshus E (2016). Variability in institutional screening practices related to collegiate student-athlete mental health. Journal of Athletic Training, 51(5), 389–397. 10.4085/1062-6050-5L5.07 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Lienert GA, & Raatz U (1998). Test Construction and Test Analysis (6th ed). Beltz. [Google Scholar]
  36. Lilenfeld LRR, Wonderlich S, Riso LP, Crosby R, & Mitchell J (2006). Eating disorders and personality: A methodological and empirical review. Clinical Psychology Review, 26(3), 299–320. 10.1016/j.cpr.2005.10.003 [DOI] [PubMed] [Google Scholar]
  37. MacCallum RC, Browne MW, & Sugawara HM (1996). Power analysis and determination of sample size for covariance structure modeling. Psychological Methods, 1(2), 130–149. [Google Scholar]
  38. Martinsen M, Holme I, Pensgaard AM, Torstveit MK, & Sundgot-Borgen J (2014). The development of the Brief Eating Disorder in Athletes Questionnaire. Medicine & Science in Sports & Exercise, 46(8), 1666–1675. 10.1249/MSS.0000000000000276 [DOI] [PubMed] [Google Scholar]
  39. McDonald AH, Pritchard M, & McGuire MK (2019). Self-reported eating disorder risk in lean and non-lean NCAA Collegiate Athletes. Eating and Weight Disorders - Studies on Anorexia, Bulimia and Obesity, 1–6. 10.1007/s40519-019-00681-0 [DOI] [PubMed] [Google Scholar]
  40. McNulty KY, Adams CH, Anderson JM, & Affenito SG (2001). Development and validation of a screening tool to identify eating disorders in female athletes. Journal of the American Dietetic Association, 101(8), 886–892. 10.1016/S0002-8223(01)00218-8 [DOI] [PubMed] [Google Scholar]
  41. Mitchell JE, & Crow S (2006). Medical complications of anorexia nervosa and bulimia nervosa. Current Opinion in Psychiatry, 19(4), 438–443. 10.1097/01.yco.0000228768.79097.3e [DOI] [PubMed] [Google Scholar]
  42. Mond JM, Hay PJ, Rodgers B, Owen C, & Beumont PJV (2004). Validity of the Eating Disorder Examination Questionnaire (EDE-Q) in screening for eating disorders in community samples. Behaviour Research and Therapy, 42(5), 551–567. 10.1016/S0005-7967(03)00161-X [DOI] [PubMed] [Google Scholar]
  43. Mountjoy M, Hutchinson M, Cruz L, & Lebrun C (2015). Female Athlete Triad Screening Questionnaire. In Female Athlete Triad Coalition. [Google Scholar]
  44. Mountjoy M, Sundgot-Borgen J, Burke L, Carter S, Constantini N, Lebrun C, Meyer N, Sherman R, Steffen K, Budgett R, & Ljungqvist A (2014). The IOC consensus statement: beyond the Female Athlete Triad--Relative Energy Deficiency in Sport (RED-S). British Journal of Sports Medicine, 48, 491–497. 10.1136/bjsports-2014-093502 [DOI] [PubMed] [Google Scholar]
  45. Murray SB, Nagata JM, Griffiths S, Calzo JP, Brown TA, Mitchison D, Blashill AJ, & Mond JM (2017). The enigma of male eating disorders: A critical review and synthesis. Clinical Psychology Review, 57, 1–11. 10.1016/j.cpr.2017.08.001 [DOI] [PubMed] [Google Scholar]
  46. Nagel DL, Black DR, Leverenz LJ, & Coster DC (2000). Evaluation of a screening test for female college athletes with eating disorders and disordered eating. Journal of Athletic Training, 35(4), 431–440. [PMC free article] [PubMed] [Google Scholar]
  47. Nunnally JC, & Bernstein IR (1994). Psychometric Theory (3rd ed). McCraw-Hill. [Google Scholar]
  48. O’Connor BP (2000). SPSS and SAS programs for determining the number of components using parallel analysis and velicer’s MAP test. Behavior Research Methods, Instruments, & Computers, 32(3), 396–402. [DOI] [PubMed] [Google Scholar]
  49. Picard CL (1999). The level of competition as a factor for the development of eating disorders in female collegiate athletes. Journal of Youth and Adolescence, 28(5), 583–594. 10.1023/A:1021606710398 [DOI] [Google Scholar]
  50. Plateau CR, Arcelus J, & Meyer C (2017). Detecting eating psychopathology in female athletes by asking about exercise: Use of the Compulsive Exercise Test. European Eating Disorders Review, 25(6), 618–624. 10.1002/erv.2561 [DOI] [PubMed] [Google Scholar]
  51. Plateau CR, Shanmugam V, Duckham RL, Goodwin H, Jowett S, Brooke-Wavell KSF, Laybourne A, Arcelus J, & Meyer C (2014). Use of the Compulsive Exercise Test with athletes: Norms and links with eating psychopathology. Journal of Applied Sport Psychology, 26(3), 287–301. 10.1080/10413200.2013.867911 [DOI] [Google Scholar]
  52. Pope Z, Gao Y, Bolter N, & Pritchard M (2015). Validity and reliability of eating disorder assessments used with athletes: A review. Journal of Sport and Health Science, 4(3), 211–221. 10.1016/jjshs.2014.05.001 [DOI] [Google Scholar]
  53. Pustivšek S, Hadžić V, Dervišević E, & Carruthers J (2019). Risk for eating disorders and body composition among adolescent female and male athletes and non-athlete controls. International Journal of Adolescent Medicine and Health. 10.1515/ijamh-2017-0190 [DOI] [PubMed] [Google Scholar]
  54. Putnick DL, & Bornstein MH (2016). Measurement invariance conventions and reporting: The state of the art and future directions for psychological research. Developmental Review, 41, 71–90. 10.1016/j.dr.2016.06.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Quick VM, & Byrd-Bredbenner C (2013). Eating Disorders Examination Questionnaire (EDE-Q): Norms for US college students. Eating and Weight Disorders, 75(1), 29–35. 10.1007/s40519-013-0015-1 [DOI] [PubMed] [Google Scholar]
  56. Rose JS, Vaewsorn A, Rosselli-Navarra F, Wilson GT, & Weissman RS (2013). Test-retest reliability of the eating disorder examination-questionnaire (EDE-Q) in a college sample. Journal of Eating Disorders, 7(1). 10.1186/2050-2974-1-42 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Schaal K, Tafflet M, Nassif H, Thibault V, Pichard C, Alcotte M, Guillet T, El Helou N, Berthelot G, Simon S, & Toussaint JF (2011). Psychological balance in high level athletes: Gender-based differences and sport-specific patterns. PLoS ONE, 6(5). 10.1371/journal.pone.0019007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Schaefer LM, Smith KE, Leonard R, Wetterneck C, Smith B, Farrell N, Riemann BC, Frederick DA, Schaumberg K, Klump KL, Anderson DA, & Thompson JK (2018). Identifying a male clinical cutoff on the Eating Disorder Examination-Questionnaire (EDE-Q). International Journal of Eating Disorders, 57(12), 1357–1360. 10.1002/eat.22972 [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Schermelleh-Engel K, & Müller H (2003). Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of Psychological Research, 5(2), 28–74. [Google Scholar]
  60. Smink FRE, van Hoeken D, Oldehinkel AJ, & Hoek HW (2014). Prevalence and severity of DSM-5 eating disorders in a community cohort of adolescents. International Journal of Eating Disorders, 47(6), 610–619. 10.1002/eat.22316 [DOI] [PubMed] [Google Scholar]
  61. Smolak L, & Murnen SK (2008). Drive for leanness: Assessment and relationship to gender, gender role and objectification. Body Image, 5(3), 251–260. 10.1016/j.bodyim.2008.03.004 [DOI] [PubMed] [Google Scholar]
  62. Song Y-M, & Lee K (2019). Eating behavior and metabolic syndrome over time. Eating and Weight Disorders. 10.1007/s40519-019-00640-9 [DOI] [PubMed] [Google Scholar]
  63. Stevens JP (1992). Applied Multivariate Statistics for the Social Sciences. (2nd ed). Lawrence Erlbaum. [Google Scholar]
  64. Stice E, Marti CN, & Rohde P (2013). Prevalence, incidence, impairment, and course of the proposed DSM-5 eating disorder diagnoses in an 8-year prospective community study of young women. Journal of Abnormal Psychology, 122(2), 445–457. 10.1037/a0030679 [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Sundgot-Borgen Jorunn Torstveit, M K. (2004). Prevalence of eating disorders in elite athletes is higher than in the general population. Clinical Journal of Sport Medicine, 14(1), 25–32. [DOI] [PubMed] [Google Scholar]
  66. Swami V, & Barron D (2019). Translation and validation of body image instruments: Challenges, good practice guidelines, and reporting recommendations for test adaptation. Body Image, 31, 204–220. 10.1016/j.bodyim.2018.08.014 [DOI] [PubMed] [Google Scholar]
  67. Tabachnick BG, & Fidell LS (2006). Using Multivariate Statistics (5th ed). Pearson. [Google Scholar]
  68. Taranis L, Touyz S, & Meyer C (2011). Disordered eating and exercise: Development and preliminary validation of the compulsive exercise test (CET). European Eating Disorders Review, 19(3), 256–268. 10.1002/erv.1108 [DOI] [PubMed] [Google Scholar]
  69. Tenforde AS, Barrack MT, Nattiv A, & Fredericson M (2016). Parallels with the Female Athlete Triad in male athletes. Sports Medicine, 46(2), 171–182. 10.1007/s40279-015-0411-y [DOI] [PubMed] [Google Scholar]
  70. Thompson JK, & Stice E (2001). Thin-ideal internalization: Mounting evidence for a new risk factor for body-image disturbance and eating pathology. Current Directions in Psychological Science, 10(5), 181–183. 10.1111/1467-8721.00144 [DOI] [Google Scholar]
  71. Udo T, & Grilo CM (2019). Psychiatric and medical correlates of DSM-5 eating disorders in a nationally representative sample of adults in the United States. International Journal of Eating Disorders, 52(1), 42–50. 10.1002/eat.23004 [DOI] [PubMed] [Google Scholar]
  72. Worthington RL, & Whittaker TA (2006). Scale development research: A content analysis and recommendations for best practices. The Counseling Psychologist, 34(6), 806–838. 10.1177/0011000006288127 [DOI] [Google Scholar]

RESOURCES