[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
November 2004

Preventing Substance Use and Disordered EatingInitial Outcomes of the ATHENA (Athletes Targeting Healthy Exercise and Nutrition Alternatives) Program

Author Affiliations

Author Affiliations: Department of Medicine, Oregon Health & Science University (OHSU), Portland (Drs Elliot, Goldberg, and Moe and Mss DeFrancesco and Durham); and the Oregon Research Institute, Eugene (Ms Hix-Small).

Arch Pediatr Adolesc Med. 2004;158(11):1043-1049. doi:10.1001/archpedi.158.11.1043

Objectives  To implement and to assess the efficacy of a school-based, sport team–centered program to prevent young female high school athletes’ disordered eating and body-shaping drug use.

Design and Setting  Prospective controlled trial in 18 high schools, with balanced random assignment by school to the intervention and usual-care control conditions.

Participants  We enrolled 928 students from 40 participating sport teams. Mean age was 15.4 years, 92.2% were white, and follow-up retention was 72%.

Intervention  The ATHENA (Athletes Targeting Healthy Exercise and Nutrition Alternative) curriculum’s 8 weekly 45-minute sessions were incorporated into a team’s usual practice activities. Content was gender-specific, peer-led, and explicitly scripted. Topics included healthy sport nutrition, effective exercise training, drug use and other unhealthy behaviors’ effects on sport performance, media images of females, and depression prevention.

Main Outcome Measures  We assessed participants by confidential questionnaire prior to and following their sport season. We determined program effects using an analysis of covariance–based approach within the Generalized Estimating Equation framework.

Results  Experimental athletes reported significantly less ongoing and new use of diet pills and less new use of athletic-enhancing substances (amphetamines, anabolic steroids, and sport supplements) (P<.05 for each). Other health-harming actions also were reduced (less riding with an alcohol-consuming driver [P = .05], more seat belt use [P<.05], and less new sexual activity [P<.05]). The ATHENA athletes had coincident positive changes in strength-training self-efficacy (P<.005) and healthy eating behaviors (P<.001). Reductions occurred in intentions toward future use of diet pills (P<.05), vomiting to lose weight (P<.05), and use of tobacco (P<.05) and muscle-building supplements (P<.005). The program’s curriculum components were altered appropriately (controlling mood [P<.005], refusal skills [P = .05], belief in the media [P<.005], and perceptions of closest friends’ body-shaping drug use [P<.001]).

Conclusions  Sport teams are effective natural vehicles for gender-specific, peer-led curricula to promote healthy lifestyles and to deter disordered eating, athletic-enhancing substance use, and other health-harming behaviors.

Approximately half of female and male high school students participate in school sports.1,2 Contrary to popular belief, these young athletes are not protected from drug use and other harmful behaviors.3,4 For young females, the sociocultural pressures toward thinness may be compounded by similar influences of their sport, resulting in more prevalent disordered eating behaviors and body-shaping drug use (tobacco, diet pills, diuretics, laxatives, amphetamines, and anabolic steroids).5,6 These health-harming actions, even among individuals lacking the full diagnostic criteria for anorexia or bulimia nervosa, can be associated with significant morbidity and mortality.7,8 Unfortunately, programs to prevent these practices largely have been unsuccessful9 or paradoxically, may have increased these detrimental behaviors.10

The high school years are important for prevention, as drug use approximately doubles,2 and middle-school prevention programs’ beneficial effects often are attenuated or lost.11 Because older adolescents’ behaviors are shaped more by peers and social influences and less by classroom activities,12 effective high school–based prevention programs are limited.13

Although sport participation may increase a young female’s risk for certain unhealthy activities, athletic teams also may provide a natural setting for bonded peers to address gender-specific topics. The sport team’s use as a vehicle for health promotion was demonstrated for male high school athletes with the ATLAS (Adolescents Training and Learning to Avoid Steroids) program, which significantly improved nutrition behaviors and exercise self-efficacy, while reducing use of alcohol, illicit drugs, and performance-enhancing supplements.14

The ATHENA (Athletes Targeting Healthy Exercise and Nutrition Alternatives) program parallels ATLAS in its school-based, team-centered format and promotion of healthy nutrition and effective exercise training as alternatives to harmful behaviors. However, ATHENA’s objectives are to reduce the young female athlete’s disordered eating habits and to deter use of body-shaping substances. To evaluate the ATHENA program’s feasibility and efficacy, we recruited 18 high schools and randomized them to an intervention and a usual-care control group. We hypothesized that program participation would result in greater positive change, which would be mediated by gender-specific influences targeted in the curriculum.


We asked 18 public high schools from northwest Oregon and southwest Washington to participate; all enrolled in the study. In selecting that number of schools, we used prevalence rates from our prior studies15,16 and assumed relatively low interdependence among students enrolled at the same school. We anticipated 18 schools would provide a greater than 90% chance of detecting intervention effects accounting for 10% or more of the variance, without covariate adjustment, for primary outcomes, such as diet pill use, disordered eating behaviors, and healthy alternative behaviors. Participating schools had a spectrum of sizes and varied locales to minimize potential cross-contamination. They were matched in balanced dyads based on size, average socioeconomic status, and student demographics. An individual unfamiliar with the participating schools used a program for generating random numbers to assign 1 member of the pair to the experimental (even) or control (odd) group, with the remaining pair member going to the other condition.

Because disordered eating is a feature among all women’s sports,6 after randomization we met with the experimental schools’ athletic directors and coaches and offered participation to all women’s sports teams, including dance and cheerleading. At least 1 or 2 teams per school chose to participate. In general, not enlisting was because of logistic and feasibility concerns, such as the coach not yet being hired or a sport’s funding being in jeopardy. For each experimental team electing to participate, we specifically recruited the similar team from its matching control school, and with that focused effort, all corresponding control teams chose to participate. Study enrollment was offered to all athletes from involved sport teams. Students and their parents or guardians provided written informed consent for study participation. The institutional review board of Oregon Health & Science University, Portland, approved the study.

Student involvement was voluntary. Study information highlighted that neither team membership nor playing time would be influenced by ATHENA participation. School personnel were not involved in recruitment, and signed consents were returned to research staff. Research assistants administered confidential questionnaires at baseline (immediately preceding the sport season) and within 2 weeks of the sport season’s conclusion. Coaches, teachers, and other school personnel were not involved with survey administration.

Survey instrument

We assessed potential influences on drug use and disordered eating, including those of peers, coaches, the media, and society.17 In addition, individuals’ knowledge and characteristics were indexed using items from our earlier research14,15 and surveys concerning disordered eating,17,18 depression,19 self-image,20 and self-esteem.21 For knowledge of specific topics, self-esteem, body image, and perceptions about peers and teammates, we grouped 3 to 7 survey items into reliable constructs (standardized α at least.78) and used the items’ mean as the score.

Acquiring new behaviors typically follows a sequence, wherein intentions toward an action increase, followed by experimenting with an activity, and finally, the behavior becomes an established practice.22,23 Higher intentions can characterize those with a greater likelihood of practicing those habits in the future. We assessed behavioral intent by asking about future actions. These and other items used a 7-point Likert scale, with anchors ranging from strongly disagree to strongly agree. Substance abuse questions used a format similar to the Monitoring the Future National Survey24 and the Youth Risk Behavior Surveillance–United States, 2001,2 with similar designs to assess supplement use and disordered eating behaviors. The final instrument contained 177 questions and could be completed in approximately 30 minutes.


The ATHENA intervention was delivered during a team’s sport season, and its eight 45-minute classroom sessions were integrated into a team’s usual practice activities. During those sessions, a team assembled as a group with students organized in coach-assigned stable learning clusters of approximately 6 students (squads). The coach also designated 1 athlete per small group as the squad leader. The squad leaders led approximately 70% of ATHENA activities, with the coach or other team staff member acting as a facilitator and timekeeper, beginning and concluding the sessions. Prior to the first classroom meeting, squad leaders attended a 90-minute orientation session. They and the coach used scripted lessons to direct the meetings. Other student group members used workbooks, and all participants received a pocket-sized sport nutrition and training guide. The guide contained reference information and was used by participants when monitoring their protein and calcium intake and for tracking their daily mood and fun activities.

The curriculum scope is based on identified features of effective prevention programs,12 sex-specific correlates of drug use and disordered eating behaviors,25,26 and our prior survey results, which defined and prioritized factors relating to higher intentions toward these health-harming practices.15,16 Each classroom session was composed of 3 to 5 activities, designed to foster active learning and application of new abilities. Topics included a balanced presentation concerning the consequences of substance use and other unhealthy behaviors and the beneficial effects of appropriate sport nutrition and effective exercise training. The nutrition content focused on classifying foods as high in carbohydrates, protein, or fats, and the need for carbohydrates as muscle “fuel” for intense exercise and achieving adequate protein and calcium intakes. Unlike the ATLAS program for young male athletes, specific caloric intake goals were not presented.

Depression is more common among females, and a low mood is a risk for a female’s disordered eating and drug use.27 To address this factor, ATHENA incorporated a depression prevention program with established efficacy,28 using sequenced cognitive restructuring tasks adapted for the sport team setting. The curriculum countered media influences by deconstructing and remaking magazine advertisements. Students also practiced relevant refusal skills, and healthy norms were reinforced as squads created and presented to their teammates public service campaigns designed to discourage drug use and disordered eating practices. In addition, at each session, participants received a specific short-term health goal, such as eating breakfast daily and recording their 24-hour calcium intake, to complete before the next meeting. The control teams were offered preprinted pamphlets concerning disordered eating, drug use, and sports nutrition.

Data analyses

We monitored implementation and fidelity using direct observation and content checklists that mirrored the scripted lesson plans. To determine baseline equivalence and intervention effects, we used an analysis of covariance–based approach within the generalized estimating-equations random-effects model framework, which extends generalized linear models to deal with time dependence and data's hierarchical structure.2931 Because the ATHENA program was designed as primary prevention, we also determined the incidence of new users (ie, athletes who began a behavior after the baseline assessment) and compared conditions using a χ2 test analysis. For those results, we report the risk ratio and its 95% confidence interval. An analogous format was applied to examine participants who stopped a behavior. Where appropriate, we adjusted the significance level for multiple comparisons using a modified Bonferroni procedure.32,33

Attrition, fidelity, and baseline characteristics

The ATHENA program was feasible and acceptable to coaches and student athletes. All enrolled teams assigned to the experimental and control conditions were retained in the study. Preseason to postseason student attrition was similar in both groups (Table 1) and comparable to the retention observed with male high school athletes.14 Unlike typical classroom programs, in addition to losses owing to study withdrawal and school transfer and drop out, attrition also included being cut from the roster, injury, and quitting the team. Fidelity to the curriculum was high, and experimental teams covered 80.8% ± 0.2% (mean ± SD) of content items per session. Table 1 also lists descriptive information by condition at the preseason assessment. Prior to the intervention, experimental and control groups had similar demographic parameters, drug use indices, knowledge scores, and self-assessed behaviors and attitudes, except for understanding of anabolic steroids. For that construct at baseline, the experimental athletes had significantly greater knowledge than the control group (P<.05) (Table 2).

Table 1. 
Baseline Characteristics*
Baseline Characteristics*
Table 2. 
Main Outcome Measures Prior to and Following the Sport Season*
Main Outcome Measures Prior to and Following the Sport Season*
Disordered eating behaviors and body-shaping substance use

Survey results prior to and within 2 weeks of the sports seasons’ conclusion indicated that following the intervention the experimental participants reported less recent (last 3 months) diet pill use (P<.05) (Table 2). The prevalence of creatine use was low, limiting the ability to demonstrate effects on use, and tobacco use decreased in both experimental and control student athletes during their sport seasons.

Although initiating drug use during the sport season was infrequent, differences between groups were observed for both diet pills and other body-shaping agents. Intervention athletes had fewer new users of diet pills (P<.05; relative risk [RR] for controls, 2.80; 95% CI, 1.02-7.68) and other body-shaping substances (amphetamines, anabolic steroids, and muscle-building supplements) (P<.05; RR for controls, 1.55; 95% CI, 1.03-2.31). After initiating the program, there also was a favorable trend for more experimental students stopping diet pills use (P = .07; RR for controls, 0.41; 95% CI, 0.15-1.13).

Nutrition, exercise abilities, and beliefs

The ATHENA experimental athletes reported significant positive changes in dietary habits and exercise training self-efficacy (Table 2). Following the intervention, experimental students indicated more tracking of their protein intake (P<.005), increased total daily protein consumption (P<.001), and greater confidence in their strength-training skills (P<.005). Accompanying those favorable changes, intervention student athletes also had significantly greater belief that what they ate affected their sport performance (P<.05).

Additional health-influencing behaviors

The groups differed in behaviors not specifically addressed in the curriculum. Following the program, experimental students were less likely to ride in a car with an alcohol-consuming driver (P = .05) and more likely to use seatbelts (P<.05). In addition, compared with experimental students more control participants became sexually active (P<.05; RR for controls, 2.50; 95% CI, 1.05-5.83). Not shown are results indicating that these favorable changes were specific to those behaviors, as no significant intervention effect was observed for “doing something against the law” and “skipping classes.” Owing to the intensity of practice and competitions, we expected recent sport injuries to be increased following the season. While that pattern was seen among control students, the experimental group’s sport injuries were reduced (P<.05), perhaps relating to the curriculum’s nutrition and physical training objectives.

Intentions toward future actions

The ATHENA program resulted in favorable changes in the experimental participants’ intentions toward future vomiting to lose weight (P<.05) and use of diet pills (P<.05), tobacco (P<.05), and creatine (P<.005). No statistically significant differences were seen in intentions toward either future or current use of alcohol and marijuana.

Curriculum content and other potential influences

Experimental participants increased their understanding and retention in the curriculum’s knowledge domains, including anabolic steroids’ effects (P<.001) and alcohol’s athletic-harming consequences (P<.001). They also had better knowledge of their calcium needs (P<.01) and the basics of sport nutrition (P<.001). Although baseline knowledge was highest for disordered eating’s consequences (ie, highest agreement with true statements), the program’s positive influence also approached significance (P = .06).

Mood, Traits, and Beliefs

Among experimental and control participants, agreement with feelings of depression was low (ie, in general, students strongly disagreed with statements characterizing depression) before and after the sport season. Despite that finding, following the ATHENA program, intervention students believed more strongly that their mood is affected by what they do (P<.01) and that they had greater skill in controlling their mood (P<.005).

The personality dimensions self-esteem and positive body image were similar for both conditions. However, personal vulnerability, indexed by disagreement with accepting the risks from drugs, differed between the groups (P<.005), with experimental students disagreeing more across the sport season. Importantly, following the intervention, curriculum participants differed from controls by indicating greater ability to turn down an offer to engage in drug use and disordered eating behaviors (P = .05).


Following the program, ATHENA athletes believed less in the truth of advertisements (P<.005) and disagreed more with a statement that men find thin women most attractive (P<.05).

Coach and Peer Effects

Perceptions about their coaches did not differ for the 2 conditions, although the observed trends among ATHENA experimental athletes were in the desired direction (Table 2). After participation, intervention students agreed less that their closest friends used body-shaping drugs (P<.001) and agreed more that their closest friends would be against their drug use (P<.001). Although perceived pressure from other females and teammates to lose weight was low, for the former, it significantly decreased among experimental student athletes (P<.05). Perceived school norms for disordered eating and body-shaping drug use did not differ between groups.


The ATHENA curriculum succeeded in most of its prevention and health promotion goals. Following their sport season, intervention students reported less ongoing and new diet pill use and less new use of athletic-enhancing, body-shaping substances (amphetamines, anabolic steroids, and muscle-building supplements). Experimental participants understood more about the presented topics, had improved self-reported dietary habits, and indicated greater self-efficacy for exercise training. They had better ability to control their mood, less belief in the media, and enhanced resistance skills. As was shown for males,14 a sport-team-centered, gender-specific, peer-led curriculum can be an effective means to alter a young female’s knowledge, attitudes, and behaviors.

Healthy changes also occurred in actions not specifically addressed during the curriculum, including riding in a car with an alcohol-consuming driver, wearing a seatbelt, and new sexual activity. Among adolescents the risk and protective factors for different health-harming behaviors frequently overlap,16,34,35 and that intersection may explain ATHENA’s extended benefits. In addition, experimental participants had reduced intentions for future disordered eating habits and body-shaping drug use. Studies have indicated that young female athletes may defer, rather than permanently avoid unhealthy habits,36 and favorably altered beliefs and skills may help protect them from future health-harming behaviors.

The ATHENA student athletes improved their self-reported nutrition habits. Cross-sectional studies document that adolescent females’ diets are far from optimal.37 The curriculum focused on understanding nutrients’ roles in sport performance; classifying common foods as mainly carbohydrate, protein, or fat; eating enough protein and calcium; and choosing appropriate preworkout and postworkout snacks. To avoid the unintended adverse effects that can accompany disordered eating prevention,10 we did not include calorie counting in the curriculum.

The ATHENA program promoted healthy alternatives to harmful actions. In a review of drug prevention programs, alternatives were the component with the largest effect size and most robust positive outcomes.38 Unlike alternatives such as attending an Outward Bound experience, which may be infeasible,38 ATHENA’s nutrition and exercise content was applicable to all students and addressed adolescent athletes’ immediate performance goals. Although not explicitly emphasized, the sport nutrition teaching also advanced components of the chronologically distant, typical school health education objectives.39

Young female athletes’ use of muscle-building agents, especially anabolic steroids, is of increasing concern.40 Adolescent females predisposed toward anabolic steroid use share characteristics with those at higher risk for unhealthy eating habits,40 suggesting that a prevention program for the latter also may deter anabolic steroid use. We found a low prevalence of anabolic steroid use, less than that in our prior study41 and that reported by others.40 Perhaps, our current investigation’s confidential (but not anonymous) questionnaire influenced athletes to underreport their anabolic steroid use. Interestingly, anabolic steroid use has been found more often among younger female adolescents,40 and our findings also could reflect our subjects being older, high school–aged teenagers.

Gender-specific risk and protective factors were emphasized in the ATHENA curriculum. Among young females, depression is related to disordered eating habits and drug use.27 Although regular exercise itself may protect from depression, among adolescents that effect seems to occur only among males.42 The ATHENA intervention’s cognitive restructuring skill-building activities, used successfully in longer programs focusing exclusively on those tasks,28 were incorporated easily and paralleled the behavioral goals and training activities typifying sport team practices.

The media depiction of idealized extremely thin, unhealthy female physiques may encourage harmful weight loss practices.5,43,44 Other health promotion programs have targeted the media.4547 However, those are designed for younger teenagers, none are school-based, and their favorable outcomes have not been confirmed with prospective randomized trials. Among other eating disorder prevention programs studied,9,10 even the most recent attempts have failed to influence behaviors, despite favorable changes in knowledge and attitudes.48

We recognize that our findings have limitations. As with any program, determining durability of change is important, and our results are relatively short-term. In addition, we found rates of substance use and unhealthy eating habits lower than those identified by national surveys.2,24 Our data are based on self-report, which can be biased by situational and cognitive factors,49 and this discrepancy may indicate underreporting by or nonparticipation of those with health-harming practices. However, when assured of confidentiality, biochemical measures generally confirm self-reports,50 and if underreporting was present, it would be expected to reduce positive program outcomes, making the ATHENA results more remarkable.

In addition, ethnic minorities were a low percentage of our study participants. This reflects both local demographics and the relatively lower participation of female minority students in school sports.51 Although recent evidence indicates risks for disordered eating are similar among ethnic groups,52,53 our findings may not be representative of minority student athletes. Finally, the ATHENA program did not influence use of alcohol and marijuana, the substances most used by adolescents.2,24 These were a minor component of the ATHENA curriculum. Longer follow-up, further emphasis on their athletic-harming effects, and reinforcement of the program’s protective skills and attitudes may extend benefits to these agents.

Sport teams can be effective vehicles to promote healthy lifestyles and to deter drug use and other harmful behaviors. The format uses influential instructors and existing, single-sex, nondeviant bonded peer groups to deliver immediately relevant content. Effects may be enhanced by continuity of coaching staff, role modeling by older athletes, and reinforcement during the many hours of team activities. Reviews of sports’ adverse consequences54 suggest athletics has moved away from its original health-enhancing goals. Incorporating peer-led drug use prevention and health promotion curricula into these settings may help rekindle and strengthen sports’ health-enhancing mission.

What This Study Adds

Disordered eating habits and body-shaping drug use are prevalent health-harming behaviors among young females, especially the 50% participating in high school athletics. Previous universal disordered eating and drug use prevention programs have changed students’ knowledge, but the influence on unhealthy behaviors has been limited. The ATHENA curriculum effectively altered targeted student abilities and reduced body-shaping drug use, while promoting healthy eating habits. As shown for male high school athletes,14 sport teams are an untapped resource for positively altering young females’ beliefs and actions.

Correspondence: Diane L. Elliot, MD, Oregon Health & Science University, CR110, 3181 SW Sam Jackson Park Rd, Portland, OR 97239-3098 (elliotd@ohsu.edu).

Financial Disclosure: Oregon Health & Science University and Drs Elliot and Goldberg have a significant financial interest from the commercial sale of technologies used in this research. This potential conflict of interest has been reviewed and managed by the OHSU Conflict of Interest in Research Committee.

Back to top
Article Information

Accepted for Publication: July 7, 2004.

Funding/Support: This study was funded by grant 5R01 DA07356 from the National Institute on Drug Abuse, National Institutes of Health, Bethesda, Md, with assistance from grant 5 M01 RR000334 from Public Health Services.

US Department of Health and Human Services, Physical Activity and Health: A Report of the Surgeon General.  Atlanta, Ga US Dept of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion1996;
Grunbaum  JAKann  LKinchen  SA  et al.  Youth Risk Behavior Surveillance—United States, 2001.  MMWR Surveill Summ 2002;511- 62PubMed
French  SAStory  MDownes  BResnick  MDBlum  RW Frequent dieting among adolescents: psychosocial and health behavior correlates.  Am J Public Health 1995;85695- 701PubMedCrossref
Naylor  AHGardner  DZaichkowsky  L Drug use patterns among high school athletes and nonathletes.  Adolescence 2001;36627- 639PubMed
Becker  AEBurwell  RAGilman  SEHerzog  DBHamburg  P Eating behaviours and attitudes following prolonged exposure to television among ethnic Fijian adolescent girls.  Br J Psychiatry 2002;180509- 514PubMedCrossref
Byrne  SMcLean  N Eating disorders in athletes: a review of the literature.  J Sci Med Sport 2001;4145- 159PubMedCrossref
Whitaker  ADavies  MShaffer  D  et al.  The struggle to be thin: a survey of anorexic and bulimic symptoms in a non-referred adolescent population.  Psychol Med 1989;19143- 163PubMedCrossref
Golden  NH Eating disorders in adolescence and their sequelae.  Best Pract Res Clin Obstet Gynaecol 2003;1757- 73PubMedCrossref
Pratt  BMWoolfenden  SR Interventions for preventing eating disorders in children and adolescents [Cochrane Review on CD-ROM CD002891].  Oxford, England Cochrane Library, Update Software2002; (2)
Carter  JCStewart  DADunn  VJFairburn  CG Primary prevention of eating disorders: might it do more harm than good?  Int J Eat Disord 1997;22167- 172PubMedCrossref
Tobler  NS Meta-Analysis of Adolescent Drug Prevention Programs: Results of the 1993 Meta-Analysis.  Bethesda, Md National Institutes of Health1997;National Institute of Drug Abuse Research Monograph, 170:5-68. Also available at:
Buddeberg-Fischer  BReel  V Preventing disturbed eating behaviors: an intervention program in Swiss high school classes.  Eat Disord 2001;9109- 124Crossref
 Making the Grade.  A Guide to School Drug Prevention Programs. Washington, DC Drug Strategies1999;
Goldberg  LMacKinnon  DElliot  DL  et al.  Preventing drug use and promoting health behaviors among adolescents: results of the ATLAS Program.  Arch Pediatr Adolesc Med 2000;154332- 338PubMedCrossref
Elliot  DGoldberg  LMoe  E  et al.  The ATHENA (Athletes Targeting Healthy Exercise and Nutrition Alternatives) program: characterizing those predisposed to drug use and eating disorders.  Med Sci Sports Exerc [abstract] 1996;28S155Crossref
Moe  ELElliot  DLDuncan  TEGoldberg  L Disordered eating and drug use: defining a model and identifying components for a prevention curriculum [abstract].  J Invest Med 2002;5094A
Neumark-Sztainer  DWall  MMStory  MPerry  CL Correlates of unhealthy weight-control behaviors among adolescents: implications for prevention programs.  Health Psychol 2003;2288- 98PubMedCrossref
Rosen  JCSilberg  NT Eating Attitudes Test and Eating Disorders Inventory: norms for adolescent girls and boys.  J Consult Clin Psychol 1988;56305- 308PubMedCrossref
Radloff  LS The CES-D scale: a self-report depression scale for research in the general population.  Appl Psychol Meas 1977;1385- 401Crossref
Offer  DHoward  KI The Offer Self-Image Questionnaire for adolescents.  Arch Gen Psychiatry 1972;27529- 537PubMedCrossref
Rosenberg  M Society and the Adolescent Self-concept.  Princeton, NJ Princeton University Press1965;
Maddahian  ENewcomb  MDBentler  PM Adolescent drug use and intention to use drugs: concurrent and longitudinal analyses of four ethnic groups.  Addict Behav 1988;13191- 195PubMedCrossref
Miller  TQ A test of alternative explanations for the stage-like progression of adolescent substance use in four national samples.  Addict Behav 1994;19287- 293PubMedCrossref
Johnston  LDO'Malley  PMBachman  JG The Monitoring the Future National Survey Results on Adolescent Drug Use: Overview of Key Findings. 2002 Bethesda, Md National Institute on Drug Abuse2002;61NIH Publication 03-5374
Tomori  MRus-Makovel  M Eating behavior, depression, and self-esteem in high school students.  J Adolesc Health 2000;26361- 367PubMedCrossref
Polivy  JHerman  CP Causes of eating disorders.  Annu Rev Psychol 2002;53187- 213PubMedCrossref
Stice  EPresnell  KSpangler  D Risk factors for binge eating onset in adolescent girls: a 2-year prospective investigation.  Health Psychol 2002;21131- 138PubMedCrossref
Clarke  GNHawkins  WMurphy  MSheeber  LBLewinsohn  PMSeeley  JR Targeted prevention of unipolar depressive disorder in an at-risk sample of high school adolescents: a randomized trial of a group cognitive intervention.  J Am Acad Child Adolesc Psychiatry 1995;34312- 321PubMedCrossref
McCullagh  P Quasi-likelihood functions.  Ann Stat 1983;1159- 67Crossref
Wedderburn  RWM Quasi-likelihood functions: generalized linear models, and the Gauss-Newton method.  Biometrika 1974;61439- 447
Zeger  SLLiang  KY The analysis of discrete and continuous longitudinal data.  Biometrics 1986;42121- 130PubMedCrossref
Legendre  PLegendre  L Numerical Ecology. 2nd English ed. Amsterdam, the Netherlands Elsevier Sci Inc1998;
Jaccard  JWan  CK LISREL approaches to interaction effects in multiple regression.  Thousand Oaks, Calif Sage Publications1996;
Neumark-Sztainer  DStory  MFrench  SA Covariations of unhealthy weight loss behaviors and other high-risk behaviors among adolescents.  Arch Pediatr Adolesc Med 1996;150304- 308PubMedCrossref
Ross  HEIvis  F Binge eating and substance abuse among male and female adolescents.  Int J Eat Disord 1999;26245- 260PubMedCrossref
Ewing  BT High school athletes and marijuana use.  J Drug Educ 1998;28147- 157PubMedCrossref
Lytle  LAKubik  MY Nutritional issues for adolescents.  Best Pract Res Clin Endocrinol Metab 2003;17177- 189PubMedCrossref
Tobler  NS Drug prevention programs can work: research findings.  J Addict Dis 1992;111- 28PubMedCrossref
National Center for Chronic Disease Prevention and Health Promotion, School health defined: a coordinated school health program. Available at:http://www.cdc.gov/nccdphp/dash/about/school_health.htmAccessed August 26, 2003
Yesalis  CEBarsukiewicz  CKKopstein  ANBahrke  MS Trends in anabolic-androgenic steroid use among adolescents.  Arch Pediatr Adolesc Med 1997;1511197- 1206PubMedCrossref
Elliot  DLMoe  ELDuncan  TGoldberg  L Who are the young women at risk for anabolic steroid use [abstract]?  Med Sci Sports Exerc 1999;31S122Crossref
Unger  JB Physical activity, participation in team sports, and risk of suicidal behavior in adolescents.  Am J Health Promot 1997;1290- 93PubMedCrossref
Andrist  LC Media images, body dissatisfaction, and disordered eating in adolescent women.  MCN Am J Matern Child Nurs 2003;28119- 123PubMedCrossref
Guillen  EOBarr  SI Nutrition, dieting, and fitness messages in a magazine for adolescent women, 1970-1990.  J Adolesc Health 1994;15464- 472PubMedCrossref
 GO GIRLS!  Program (Giving Our Girls Inspiration and Resources for Lasting Self-esteem) Available at:http://www.nationaleatingdisorders.org/p.asp?WebPage_ID=296Accessed August 26, 2003
US Department of Health and Human Services, Girl Power. Available at:www.girlpower.govAccessed August 26, 2003
Frank  M GoGirlGo project survey analysis (year one). Available at:http://www.womenssportsfoundation.org/cgi-bin/iowa/sports/ggg/article.html?record=55Accessed August 26, 2003
Steiner-Adair  CSjostrom  LFranko  DL  et al.  Primary prevention of risk factors for eating disorders in adolescent girls: learning from practice.  Int J Eat Disord 2002;32401- 411PubMedCrossref
Brener  NDJohn  OGGrady  WR Assessment of factors affecting the validity of self-reported health-risk behavior among adolescents: evidence from the scientific literature.  J Adolesc Health 2003;33436- 457PubMedCrossref
Forman  SGLinney  JA Increasing the validity of self-report data in effectiveness trials. Leakefeld  CGBukoski  WJeds. National Institute of Drug Abuse Research Monograph 1991;107235- 247
Lopiano  D Gender equity and the black female in sport. Available at:http://www.womenssportsfoundation.org/cgi-bin/iowa/issues/disc/article.html?record=869Accessed August 26, 2003
Perez  MJoiner  TE  Jr Body image dissatisfaction and disordered eating in black and white women.  Int J Eat Disord 2003;33342- 350PubMedCrossref
Tsai  GCurbow  BHeinberg  L Sociocultural and developmental influences on body dissatisfaction and disordered eating attitudes and behaviors of Asian women.  J Nerv Ment Dis 2003;191309- 318PubMed
Miracle  AWRees  R Lessons of the Locker Room: The Myth of School Sports.  Buffalo, NY Prometheus Books1994;