Abstract
The Eating Attitudes Test (EAT-40), originally conceptualized and validated by David M. Garner and Paul E. Garfinkel in 1979 at the Clarke Institute of Psychiatry, University of Toronto, stands as one of the most widely utilized and foundational psychometric screening instruments in the field of eating disorder research and clinical assessment. Developed primarily to provide an objective, standardized index of the multidimensional cognitive, affective, and behavioral symptoms characteristic of anorexia nervosa, the instrument comprises 40 self-report items evaluated on a 6-point forced-choice response format. Psychometric evaluations across diverse clinical and non-clinical cohorts consistently demonstrate robust internal consistency, with Cronbach's alpha coefficients typically ranging between .79 and .94 in clinical samples and exceeding .80 in community populations. Factor analytic investigations historically yield multidimensional structural models—most notably delineating domains such as restrictive dieting, food preoccupation, bulimic behaviors, perceived social pressure to gain weight, and oral self-control. Utilizing an atypical clinical scoring system (3-2-1-0-0-0), a cumulative cutoff score of 30 reliably discriminates individuals meeting formal diagnostic criteria for anorexia nervosa and associated eating disturbances from non-clinical controls, demonstrating high diagnostic sensitivity (frequently exceeding 85%) and specificity. This article provides an exhaustive, critical psychometric review of the EAT-40, detailing its theoretical underpinnings, structural validity, reliability parameters, clinical and cross-cultural utility, scoring algorithms, and full scale inventory.
Keywords
Eating Attitudes Test, EAT-40, anorexia nervosa, bulimia nervosa, eating disorders, psychometrics, screening instrument, body image, dieting behavior, Garner and Garfinkel, food preoccupation, oral control, clinical assessment, criterion validity, factor structure
Authors
The Eating Attitudes Test (EAT-40) was developed through the collaborative academic and clinical investigation of two prominent figures in psychiatric research:
- David M. Garner, Ph.D.: A distinguished clinical psychologist whose pioneering work at the Clarke Institute of Psychiatry (now part of the Centre for Addiction and Mental Health – CAMH) and the Department of Psychiatry at the University of Toronto revolutionized the psychometric assessment of eating pathology. Dr. Garner subsequently served as the Director of the River Centre Clinic in Ohio, publishing extensively on the diagnosis, etiology, and cognitive-behavioral treatment of eating disorders, and authoring renowned diagnostic inventories including the Eating Disorder Inventory (EDI).
- Paul E. Garfinkel, M.D., FRCPC: A globally recognized psychiatrist, Professor Emeritus in the Department of Psychiatry at the University of Toronto, and former President and CEO of the Centre for Addiction and Mental Health. Dr. Garfinkel's extensive neurobiological, epidemiological, and clinical investigations have shaped modern conceptualizations of psychosomatic medicine, neuroendocrinology, and diagnostic classification systems for anorexia nervosa and bulimia nervosa.
Correspondence regarding the historical development and foundational validation studies was coordinated through the Clarke Institute of Psychiatry, 250 College Street, Toronto, Ontario, Canada.
Purpose
The primary clinical and research purpose of the EAT-40 is to quantify a broad spectrum of overt behaviors, attitudes, and cognitive distortions that typify the psychopathology of anorexia nervosa and related eating disorders. Prior to the instrument's introduction in 1979, the psychiatric assessment of anorexia nervosa relied predominantly on unstructured clinical interviews, subjective observational notes, and projective testing techniques. These modalities lacked standardized cross-cohort reliability, possessed dubious psychometric validity, and were ill-suited for large-scale epidemiological screening or quantitative longitudinal outcome monitoring.
The EAT-40 was designed to address these methodological deficits by serving four critical functions across research and clinical domains:
- Standardized Clinical Screening: Operating as a front-line, rapid screening battery capable of identifying individuals at heightened risk for eating disorders in secondary schools, universities, athletics programs, modeling agencies, and general community health contexts.
- Differential Severity Quantification: Providing a continuous, objective metric that reflects the severity of eating-related psychopathology, enabling clinicians to monitor treatment response, cognitive recovery, and behavioral symptom remission across inpatient, day-treatment, and outpatient cognitive-behavioral interventions.
- Epidemiological Case Finding: Supplying epidemiologists and psychiatric researchers with a cross-culturally replicable survey instrument to estimate prevalence, incidence, and demographic risk factors associated with subthreshold and clinical eating disturbances.
- Research Phenotyping: Serving as an empirical phenotype descriptor in biological, genetic, and neuroimaging investigations exploring the cognitive, neurochemical, and somatic correlates of extreme dietary restriction, body image disturbance, and food preoccupation.
Importantly, Garner and Garfinkel explicitly designed the EAT-40 not as an isolated, definitive diagnostic instrument, but rather as an economical, psychometrically rigorous screening threshold tool. A score falling at or above the designated clinical cutoff signifies marked eating pathology requiring comprehensive, multidimensional psychiatric and nutritional evaluation.
Psychological Construct
The EAT-40 operationalizes the complex, multifaceted construct of eating-related psychopathology. Rather than viewing eating pathology as a uniform or unidimensional behavioral deviance, the scale encapsulates affective, behavioral, cognitive, and physiological dimensions that together constitute the clinical gestalt of anorexia nervosa and bulimia nervosa. These dimensional facets encompass:
1. Restrictive Dieting and Caloric Preoccupation
This core dimension measures intentional dietary deprivation driven by an intense, persistent fear of weight gain. Cognitively, it involves extreme awareness of caloric, carbohydrate, and nutrient density, pervasive monitoring of food intake, and the rigid moral categorization of nutritional substances into binary "safe" versus "forbidden" groups. Behaviors include the systematic avoidance of high-carbohydrate or sugary foods, reliance on diet products, and rituals such as cutting food into minute morsels to delay ingestion.
2. Bulimic Impulses and Compensatory Behaviors
Although initially developed to measure anorexia nervosa, the EAT-40 captures severe bulimic episodes and purging symptomatology. It examines the loss of perceived self-control during food consumption (bingeing), pervasive postprandial guilt, emotional distress, and compensatory purging behaviors such as self-induced vomiting, laxative abuse, and strenuous, compulsive exercise aimed at offsetting caloric consumption.
3. Food Preoccupation and Intrusive Cognition
This dimension operationalizes the obsessional nature of starvation biology. As observed in classic human starvation experiments (e.g., the Minnesota Starvation Experiment), prolonged caloric deficit generates continuous, intrusive cognitive processing focused on culinary preparations, recipe collecting, and cooking for others while rigidly refusing personal consumption. The construct captures the perception that thoughts of food dominate daily cognitive capacity.
4. Perceived External Pressure and Social Discomfort Around Eating
This component taps interpersonal and social dimensions of eating disturbances. It measures interpersonal distress triggered by communal dining, heightened sensitivity to external scrutiny, avoidance of social meals or restaurants, and the pervasive paranoid or resistant belief that family members, peers, or clinicians are coercively attempting to force weight restoration.
5. Drive for Thinness and Body Shape Preoccupation
Central to modern diagnostic formulations in the DSM-5-TR, this dimension encapsulates extreme body dissatisfaction, weight phobia (am terrified about being overweight), persistent cognitive preoccupation with adipose tissue, preference for an empty stomach, and frequent, compulsive self-weighing rituals designed to validate self-worth via somatic metrics.
6. Oral Self-Control and Asceticism
This construct assesses the rigid internal mechanisms individuals employ to repress physiological hunger cues. It represents an ascetic valuation of absolute self-denial, wherein experiencing hunger, ignoring bodily appetitive signals, and actively enforcing somatic mastery provide intense affective reinforcement and a paradoxical sense of personal empowerment.
Theoretical Framework
The development of the EAT-40 is grounded in several converging psychological and psychiatric paradigms that emerged during the late 20th century:
The Cognitive-Behavioral Formulation of Eating Disorders
Foremost among the conceptual frameworks informing the EAT-40 is the cognitive-behavioral model, later comprehensively synthesized by Christopher G. Fairburn. Under this theoretical rubric, the central organizing pathology of eating disorders is the over-evaluation of shape, weight, and their control. Unlike individuals without eating pathology, who evaluate self-worth across varied life domains (e.g., interpersonal relationships, professional achievements, hobbies), individuals with eating disorders judge self-worth almost exclusively in terms of their body weight, physical dimensions, and capacity to maintain rigid control over food consumption. The cognitive items of the EAT-40 directly assess these cognitive schema and the self-perpetuating maintenance loops linking restrictive dieting to physiological hunger, affective vulnerability, and compensatory purging.
Hilde Bruch's Psychosomatic and Perceptual Model
The foundational clinical conceptualizations of Hilde Bruch (1973) profoundly influenced Garner and Garfinkel's scale design. Bruch identified three primary pathognomonic characteristics in primary anorexia nervosa: a severe disturbance in body image reaching delusional proportions; an inability to correctly identify internal visceral sensations (visceroceptive deficits, including hunger, satiety, and affective states); and a paralyzing sense of pervasive ineffectiveness. Items on the EAT-40 assessing the preference for an empty stomach, bloating after meals, and viewing food as an external governing entity reflect Bruch's emphasis on visceral misperception and the desperate quest for personal agency through somatic regulation.
The Sociocultural Tripartite Influence Model
The sociocultural paradigm posits that eating disturbances arise within Westernized societies characterized by an intense cultural glorification of thinness and stigmatization of adiposity. Sociocultural pressures from peers, family, and media foster the internalization of an unrealistically slender body ideal. The EAT-40 operationalizes this internalized cultural pressure through items evaluating explicit terror of weight gain, rigid avoidance of carbohydrate-rich food groups, and comparative behaviors regarding food consumption in interpersonal environments.
Validity
Extensive psychometric investigations have established the construct, criterion, convergent, and discriminant validity of the EAT-40 across both clinical and epidemiological populations.
Criterion-Related and Diagnostic Validity
In the seminal validation study by Garner and Garfinkel (1979), the EAT-40 was administered to an initial cohort of 33 female patients meeting rigorous diagnostic criteria for anorexia nervosa and 59 female normal control subjects. The clinical anorexia nervosa cohort exhibited a mean total score of 58.9 (SD = 13.3), whereas the normal control group demonstrated a mean score of 15.6 (SD = 9.3). This intergroup disparity was statistically significant at p < .001, demonstrating profound discriminative power. Receiver operating characteristic (ROC) evaluations confirm that utilizing a clinical cutoff score of 30 yields diagnostic sensitivity typically exceeding 85% and specificity approaching 90%, with low false-negative classifications in actively restricting cohorts.
Convergent and Concurrent Validity
The concurrent validity of the EAT-40 has been repeatedly confirmed through statistically significant positive correlations with corresponding clinical and self-report measures of eating pathology, body image disturbance, and affective distress:
- Eating Disorder Inventory (EDI): EAT-40 scores demonstrate strong positive correlations (r = .70 to .85) with the Drive for Thinness, Bulimia, and Body Dissatisfaction subscales of the EDI.
- Bulimic Investigatory Test, Edinburgh (BITE): Moderate-to-high correlations (r = .60 to .75) are observed between the EAT-40 total score and BITE symptom severity scores.
- Beck Depression Inventory (BDI): Given the clinical comorbidity between major depressive disorders and eating pathology, EAT-40 total scores exhibit moderate, theoretically expected correlations (r = .40 to .55) with the BDI, indicating shared negative affectivity without construct conflation.
Discriminant Validity
Discriminant validity is confirmed by the scale's capacity to differentiate patients with anorexia nervosa from other clinical psychiatric populations, including individuals with primary unipolar depression, obsessive-compulsive disorder, and generalized anxiety disorder who may exhibit secondary appetite loss without characteristic cognitive shape and weight preoccupations. Studies show that depressed non-eating-disordered individuals score significantly below the EAT-40 cutoff of 30, confirming that the tool specifically measures eating disorder psychopathology rather than general psychiatric distress or somatic malaise.
Reliability
The reliability parameters of the EAT-40 have undergone extensive evaluation across diverse demographic settings, language adaptations, and clinical cohorts.
Internal Consistency
In the original validation cohort, Garner and Garfinkel (1979) reported an internal consistency coefficient (Cronbach's alpha) of .79 for the combined clinical anorexia nervosa cohort and .94 when calculated across the pooled clinical and control validation sample. Subsequent large-scale clinical validation trials have replicated these findings, yielding alpha coefficients typically fluctuating between .85 and .93 in adolescent and adult clinical cohorts, and between .80 and .89 in unselected collegiate and adolescent female community samples. These coefficients indicate high item homogeneity and consistent measurement of the underlying construct.
Test-Retest Reliability and Temporal Stability
Temporal stability assessments over short- to medium-term intervals demonstrate strong test-retest reliability:
- Across a 2- to 3-week test-retest interval in non-clinical collegiate cohorts, stability coefficients range from r = .84 to .89, demonstrating excellent measurement consistency in the absence of clinical intervention.
- Over longer durations (e.g., 6 to 12 months) in longitudinal adolescent community studies, test-retest correlations remain moderate (r = .65 to .72), reflecting natural developmental fluctuations in dieting behavior and body image distress.
Additionally, split-half reliability analyses conducted in early psychometric studies demonstrated Guttman split-half coefficients exceeding .85, further supporting the structural stability of the instrument.
Factor Analysis
The structural dimensionality of the EAT-40 has been investigated through both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), uncovering nuanced latent constructs that ultimately informed subsequent scale refinements in psychometrics.
Original Exploratory Factor Analysis (Garner & Garfinkel, 1979)
In their initial factor extraction utilizing principal component analysis with varimax rotation, Garner and Garfinkel extracted multiple latent factors explaining significant portions of common variance. However, many items demonstrated cross-loadings or primary affiliations with clinical features specific to restricting anorexia nervosa. A subsequent milestone factor analytic study conducted by Garner, Olmsted, Bohr, and Garfinkel (1982) on a combined sample of 160 anorexia nervosa patients revealed that a reduced set of 26 items clustered cleanly into three robust, clinically interpretable factors, giving rise to the abbreviated EAT-26:
- Factor I: Dieting (13 items): Defined by items reflecting intense avoidance of fattening foods, calorie awareness, and preoccupation with thinness (e.g., items 10, 15, 29).
- Factor II: Bulimia and Food Preoccupation (6 items): Characterized by items tapping loss-of-control bingeing, self-induced vomiting, and intrusive ruminations regarding culinary preparations (e.g., items 6, 7, 13, 40).
- Factor III: Oral Control (7 items): Comprising items relating to self-imposed restraint, avoidance of eating in response to hunger cues, and external perception of social pressure to gain weight (e.g., items 5, 8, 12, 32).
Multidimensional Structural Models in the Full 40-Item Instrument
When evaluated in its full 40-item presentation across diverse international samples, researchers (such as Wells et al., Carter & Moss) have documented more complex factor solutions, ranging from 5-factor to 7-factor models:
- Dietary Restriction and Calorie Awareness
- Food Preoccupation and Asceticism
- Purging and Bulimic Symptoms
- Social Pressure and Interpersonal Dining Distress
- Oral Self-Discipline
- Somatic Symptoms (bloating, constipation)
- Amenorrhea and Endocrine/Physiological Correlates
Modern confirmatory factor analytic (CFA) investigations often report adequate fit indices (e.g., Comparative Fit Index [CFI] > .90, Root Mean Square Error of Approximation [RMSEA] < .06) for higher-order bifactor models or modified multidimensional structures, although the presence of low-loading somatic items (e.g., constipation, regular menses) has led psychometricians to note that the abbreviated EAT-26 provides a more parsimonious factor structure for general community screening.
Instrument / Measurement Tool
- Instrument Name: Eating Attitudes Test (EAT-40)
- Original Authors: David M. Garner, Ph.D., and Paul E. Garfinkel, M.D. (1979)
- Target Population: Adolescents and adults (typically ages 12 years and older); widely utilized across clinical psychiatric settings, university student populations, athletic teams, and community cohorts.
- Administration Format: Self-report paper-and-pencil questionnaire or computerized psychological assessment battery.
- Administration Time: Approximately 7 to 10 minutes.
- Item Count: 40 items (clinical diagnostic and behavioral symptom items).
- Response Scale: 6-point forced-choice scale:
- 1 = Always
- 2 = Very Often
- 3 = Often
- 4 = Sometimes
- 5 = Rarely
- 6 = Never
- Scoring Algorithm:
- For standard (symptomatic) items, responses are recoded as follows: Always = 3, Very Often = 2, Often = 1, and the three remaining responses (Sometimes = 0, Rarely = 0, Never = 0) are assigned a score of zero.
- For reversed (asymptomatic/healthy) items (such as items 1, 18, 19, 23, 26, 38 in full original formulations), scoring is inverted: Never = 3, Rarely = 2, Sometimes = 1, and Always = 0, Very Often = 0, Often = 0.
- Total scores are derived by summing all recoded item scores, yielding a theoretical total score ranging from 0 to 120 (for a 40-item inventory) or 0 to 117 (if a 39-item enumeration is applied).
- Clinical Cutoff Score: A cumulative score of 30 or higher indicates significant eating disorder pathology, elevated risk for anorexia nervosa or bulimia nervosa, and the necessity of immediate, in-depth diagnostic clinical evaluation.
Permissions & Fee and Test Year
The Eating Attitudes Test (EAT-40) was initially published in 1979 in the journal Psychological Medicine. Under standard psychometric conventions and the authorial intent of Dr. David M. Garner, the instrument is generally accessible for non-profit academic research, clinical screening, and educational applications without the payment of licensing fees, provided that appropriate bibliographic attribution is maintained.
Commercial utilization, pharmaceutical clinical trials, integration into proprietary digital health platforms, or fee-for-service applications require explicit permission from the copyright holders or authorized representatives. Interested researchers and clinicians may consult Dr. David M. Garner's official informational resources or academic archives to verify current usage guidelines and obtain permission documentation where required.
References
- Bruch, H. (1973). Eating disorders: Obesity, anorexia nervosa, and the person within. Basic Books. https://doi.org/10.1080/00029157.1974.10403698
- Fairburn, C. G., Cooper, Z., & Shafran, R. (2003). Cognitive behaviour therapy for eating disorders: A "transdiagnostic" theory and treatment. Behaviour Research and Therapy, 41(5), 509–528. https://doi.org/10.1016/S0005-7967(02)00088-8
- Garner, D. M., & Garfinkel, P. E. (1979). The Eating Attitudes Test: An index of the symptoms of anorexia nervosa. Psychological Medicine, 9(2), 273–279. https://doi.org/10.1017/s0033291700030762
- Garner, D. M., Olmsted, M. P., Bohr, Y., & Garfinkel, P. E. (1982). The Eating Attitudes Test: Psychometric features and clinical correlates. Psychological Medicine, 12(4), 871–878. https://doi.org/10.1017/s0033291700049163
- Garner, D. M., Olmsted, M. P., & Polivy, J. (1983). Development and validation of a multidimensional eating disorder inventory for anorexia nervosa and bulimia. International Journal of Eating Disorders, 2(2), 15–34. https://doi.org/10.1192/bjp.150.1.18
- Stice, E. (2002). Risk and maintenance factors for eating pathology: A meta-analytic review. Psychological Bulletin, 128(5), 825–848. https://doi.org/10.1037/0033-2909.128.5.825