EAT-26 (Eating Attitudes Test): Take It, Score It, Understand Your Results
The EAT-26 (Eating Attitudes Test) is a brief self-report screening tool for attitudes and behaviors associated with eating disorders. Developed by Garner, Olmsted, Bohr, and Garfinkel in 1982 as an abbreviated 26-item version of the original EAT-40, it is the most widely used standardized eating disorder screen worldwide, used in schools, primary care, college counseling, athletic programs, and research. The EAT-26 covers 26 items across three subscales, takes about 5 to 10 minutes to complete, and uses a cutoff of 20 or higher to recommend further professional evaluation. The EAT-26 is a screening tool, not a diagnosis — a score at or above 20 means a qualified clinician should conduct an in-person interview, and a score below 20 does not rule out a serious eating problem. For eating-disorder-specific support, the NEDA Helpline is 1-800-931-2237.
How to interpret your EAT-26 score
The EAT-26 is scored item by item, with the total score and a single cutoff doing most of the interpretive work. The instrument author’s site is explicit that a score of 20 or higher is the threshold for professional referral. A score at or above the cutoff means the person has reported a level of eating-related concern that warrants an in-person interview with a clinician — not that the person has an eating disorder.
Items 1 through 25 use a six-point response scale from “Always” to “Never.” Only the top three options carry weight; the bottom three score zero. Item 26 is reverse-scored. That single item flips the directionality of its scale.
The item scoring rules at a glance
| Response | Items 1-25 | Item 26 (reverse-scored) |
|---|---|---|
| Always | 3 | 0 |
| Usually | 2 | 0 |
| Often | 1 | 0 |
| Sometimes | 0 | 1 |
| Rarely | 0 | 2 |
| Never | 0 | 3 |
Summing across all 26 items produces a total score from 0 to 78. The instrument author specifies that “individuals who score 20 or more on the test should be interviewed by a qualified professional to determine if they meet the diagnostic criteria for an eating disorder”. Missing-data guidance allows interpolation using the median subscale value, up to one missing value per subscale.
What a score below 20 does and does not mean
This caveat runs in both directions. The instrument author is direct: “high scores do not always reflect over-concern about body weight, body shape, and eating. Screening studies have shown that some people with high scores do not have eating disorders”. The reverse caveat is equally explicit: “If you have a low score on the EAT-26 (below 20), you still could have a serious eating problem”. The EAT-26 is a screening tool, not a diagnosis. A number below the cutoff is reason to reflect on whether other warning signs are present, not reason to dismiss concern.
The three referral criteria: score, behaviors, and body measures
The total score is one input into a referral decision, not the entire decision. The instrument author specifies three criteria. They are: “1) the total score based on the answers to the EAT-26 questions; 2) answers to the behavioral questions related to eating symptoms and weight loss, and 3) the individual’s body mass index (BMI) calculated from their height and weight”. Each part adds information the others cannot supply on their own.
Part A: the 26 symptom and attitude items
Part A is the EAT-26 proper — the 26 items that generate the total score. The factor analysis underlying the instrument produced three subscales that capture different facets of disordered-eating attitudes. This is where the cutoff of 20 applies, and a high score in Part A on its own is enough to trigger a referral.
Part B: behavioral follow-up items
Part B asks structured questions about behaviors associated with eating disorders, including weight-loss attempts and compensatory behaviors. These items are answered separately from the 26 symptom items and do not contribute to the cutoff-of-20 total. A person can score below 20 in Part A and still trigger a referral if Part B reports clinically significant behaviors.
Part C: body measurements
Part C uses body measurements as a third referral input. A clinician uses these measurements together with the score and the behavioral answers to decide whether to recommend further evaluation. The combination of inputs is what allows the EAT-26 to flag concerns that would slip past any single criterion. No single number from any one part of the assessment is a diagnosis — the EAT-26 remains a screening tool, not a diagnostic instrument.
What the EAT-26 cannot tell you: the limits of a screening tool
A screening tool is built to flag people for more thorough evaluation, not to substitute for it. The instrument author states this plainly. The EAT-26 is “not designed to make a diagnosis of an eating disorder or to take the place of a professional diagnosis or consultation. The EAT-26 alone does not yield a specific diagnosis of an eating disorder”.
Screen versus diagnosis
A diagnosis requires a clinical interview against DSM-5-TR criteria, often combined with a physical exam and laboratory testing. A clinician can rule conditions in or out, gather episode history that yes/no items cannot capture, and consider co-occurring conditions. NIMH notes that people with eating disorders frequently have co-occurring depression, anxiety, and substance use disorders, which compound severity.
False positives and false negatives both happen
Wikipedia notes “high false-positive rates and low predictive power for screening for AN and bulimia nervosa (BN) in non-clinical settings”. The instrument author echoes the false-positive direction: “high scores do not always reflect over-concern”. The false-negative direction is equally explicit: a low score below 20 does not rule out a serious eating problem.
Conditions the EAT-26 was not designed to catch
The EAT-26 was originally validated in 160 female anorexia nervosa patients and 140 female comparison subjects. Its evidence base is strongest for anorexia-spectrum and bulimia-spectrum symptoms. Several conditions sit outside that range:
- Binge eating disorder (BED). Wikipedia notes the EAT-26 “performs poorly for binge eating disorder and other specified eating disorders”. BED is the most prevalent eating disorder in the U.S., so a negative EAT-26 in someone with binge-eating concerns is not reassuring on its own.
- ARFID, pica, and rumination disorder. Cleveland Clinic’s six-type taxonomy includes ARFID (Avoidant/Restrictive Food Intake Disorder), pica, and rumination disorder alongside anorexia nervosa, bulimia nervosa, and BED. These are not part of what the EAT-26 was built to detect.
Modern measurement concerns
A 2022 Rasch analysis of the EAT-26 in 469 adults identified measurement issues. Seven items were flagged as poor fit, and the six-category Likert scale “did not function well”. The authors flagged “several concerns” with the psychometric evaluation of the EAT-26. They questioned its utility for “assessing ED risk in individuals at low risk for ED,” especially in samples of people with overweight and obesity seeking weight loss treatment. Wikipedia adds that EAT-26 stability is “moderate over two years, but vulnerable to fluctuations over four years”. The EAT-26 remains a screening tool, not a diagnosis.
EAT-26 vs other eating disorder screens
The EAT-26 is the most widely used self-report eating disorder screen, but it is not the only one, and it is not the right tool for every population. A clinician choosing a screen weighs how brief it needs to be, which populations it is validated in, and which conditions it is designed to detect.
How the EAT-26 fits into the broader screening landscape
| Tool | What it screens for | Where it fits |
|---|---|---|
| EAT-26 | Anorexia-spectrum and bulimia-spectrum symptoms in adolescents and adults | The most widely used self-report eating disorder screen worldwide |
| EAT-40 | Same domain, with 14 additional items | EAT-26 correlates r = 0.98 with the EAT-40 and is generally preferred for brevity |
| BED-focused screens | Binge eating patterns specifically | A complement when BED is suspected — the EAT-26 performs poorly for BED |
Why pairing screens matters
People with eating disorders frequently have co-occurring depression, anxiety, and substance use disorders. A clinician evaluating someone who screens positive on the EAT-26 will often look at mood and anxiety as well. Symptomatik’s mental health screens are built to pair with eating disorder tools for exactly this reason. The PHQ-9 covers depression, and the GAD-7 covers anxiety. No screen on its own is a diagnosis.
Choosing a screen by population
The EAT-26 was validated in adolescent and adult female samples. Wikipedia notes its limitations in non-clinical settings and in groups outside the original validation population. For younger children, child-specific eating attitudes assessments exist. The 2022 Rasch analysis raised specific concerns about EAT-26 performance in samples with elevated body weight. The right screen depends on the person being assessed.
When to talk to a clinician
A positive EAT-26 result — a total score of 20 or higher, concerning behavioral responses in Part B, or concerning body measurements in Part C — is a clear signal to schedule a clinical evaluation. A negative result paired with persistent concerns is the same signal. A low score below 20 does not rule out a serious eating problem. The Office on Women’s Health is direct: “All eating disorders are dangerous if left untreated”.
Consider scheduling an evaluation in any of these situations:
- You scored 20 or higher on the EAT-26. Individuals at or above this cutoff “should be interviewed by a qualified professional to determine if they meet the diagnostic criteria for an eating disorder”.
- You scored below 20 but have ongoing eating-related concerns. Persistent worry about food, body, or eating behavior is reason enough to talk to a clinician.
- You have warning signs regardless of score. Eating disorders are serious medical conditions, not a lifestyle choice. Body or weight changes can be subtle or absent in serious illness.
- You suspect binge eating disorder. The EAT-26 performs poorly for BED, and BED is the most prevalent eating disorder in the U.S..
- You are supporting someone else who screened positive. Sharing the result with a primary care clinician, mental health professional, or eating disorder specialist is a constructive next step. NIMH notes that eating disorders “can be treated successfully,” with early intervention being critical.
What to expect at a clinical evaluation
A clinician typically conducts a structured interview against DSM-5-TR criteria, performs a physical exam, and may order laboratory tests depending on what the history surfaces. The care team may include a primary care clinician, a mental health professional, and a registered dietitian. Bringing the completed EAT-26 gives the clinician a useful starting point.
Crisis and support resources
For eating-disorder-specific support: NEDA Helpline: 1-800-931-2237 and NEDA Crisis Text Line: text “NEDA” to 741741 for 24/7 crisis text support. The Alliance for Eating Disorders Awareness (ANAD) offers a peer-support helpline at (888) 375-7767, Monday through Friday from 9 a.m. to 9 p.m. CST; ANAD provides treatment referrals and peer encouragement but is not a crisis or medical service.
For any safety concern, suicidal ideation, or self-harm risk, call or text 988 to reach the U.S. Suicide & Crisis Lifeline. For a medical emergency — severe dehydration, fainting, chest pain, or any life-threatening situation — call 911 or go to the nearest emergency room.
Frequently asked questions
Where can I get the official EAT-26?
The EAT-26 is hosted on the instrument author’s own site, which provides the items, scoring rules, and the three-part referral framework. Patient-facing organizations including NEDA also offer the EAT-26 as a free, confidential screening tool.
What age range is the EAT-26 designed for?
The 1982 Garner validation enrolled adolescent and adult female samples — 160 women with anorexia nervosa and 140 female comparison subjects. The EAT-26 is most commonly used in adolescents and adults. Younger children typically need a different screen, often combined with parent input.
Is the EAT-26 valid for men?
The original validation population was female, and the published evidence in men is thinner. Wikipedia notes the EAT-26 shows “high false-positive rates and low predictive power” in non-clinical settings. Men can take the EAT-26, but the result is best interpreted by a clinician familiar with the validation populations.
Can the EAT-26 diagnose binge eating disorder?
No. Wikipedia is explicit that the EAT-26 “performs poorly for binge eating disorder and other specified eating disorders”. BED is the most prevalent eating disorder in the U.S., and a person with binge-eating concerns needs a different screen rather than relying on a low EAT-26 score for reassurance.
How long does the EAT-26 take?
The EAT-26 is a brief self-report instrument and takes only a few minutes to complete. It is “designed to be administered by mental health professionals, school counselors, coaches, camp counselors, and others with interest in gathering information”. It works in school, athletic, primary care, and outpatient psychiatric settings.
What if I scored below 20 but I am worried?
A low score does not rule out a serious eating problem. The instrument author states that “if you have a low score on the EAT-26 (below 20), you still could have a serious eating problem”. Persistent worry is a reason to talk to a clinician regardless of the number. NIMH adds that eating disorders “can be treated successfully” when detected early.
Can I retake the EAT-26 to track progress?
The EAT-26 is best used as a periodic screen, not a frequent tracker. Wikipedia notes that stability is “moderate over two years, but vulnerable to fluctuations over four years”. The 2022 Rasch analysis added measurement concerns at the item level. Clinicians evaluating change typically combine additional tools and observation rather than rely on repeated EAT-26 scores alone.