EAT-26 Eating Attitudes Test

What EAT-26 measures

EAT-26 is a 26-item attitudes screen for eating disorders in adults. Built in 1979-1982 by David Garner and colleagues at Toronto General Hospital, the original 40-item version was shortened to 26 items in Garner 1982 (Psychological Medicine 12:871-878, PMID 6961471) and has been one of the most widely used eating-disorder screens worldwide for the last four decades.

The 26 items cover three subscales: Dieting (the largest, capturing fear of weight gain, calorie awareness, food restriction), Bulimia & food preoccupation (binge episodes, mental space taken up by food, urge to purge), and Oral control (eating in front of others, self-control around food, social pressure around eating). Each item is rated on a 6-point Likert scale from Always to Never. Items 1-25 score points only for the three highest-symptom answers (Always = 3, Usually = 2, Often = 1, the rest = 0); item 26 is reverse-scored. Total range: 0-78.

The cut-off of 20 was chosen by Garner because it captured almost all of the clinical samples in development without flagging most of the non-clinical sample. The 20+ band has 70-80% positive predictive value for an eating disorder diagnosis in non-clinical screening populations.

How your score is calculated

EAT-26 is an attitudes screen, not a behavioural screen. The published EAT-26 protocol asks five behavioural follow-up questions separately (binge episodes, self-induced vomiting, laxative or diuretic use for weight control, exercise to compensate for eating, weight loss greater than 9 kg in 6 months). Any of those behaviours warrants a clinical conversation regardless of the score.

What your score means

A score of 0-19 is the band Garner 1982 saw in most adults without an eating disorder, including ones who watched their weight or had passing thoughts about food. A low score does not rule out a problem — people with binge eating disorder, atypical anorexia or sub-clinical struggles can still score under 20. The single most useful question to ask yourself: does food, weight or shape take up so much mental space that it crowds out other things? If yes, the score is less reassuring than it looks.

A score of 20-78 sits at or above the cutoff. The pattern of attitudes you reported is closer to what Garner saw in clinical samples than to the general population. This is a screen, not a diagnosis — but a conversation with a clinician within the next few weeks is the right next step. Roughly 5-10% of the general adult population scores 20+ at any time; many of them are in active or sub-clinical struggles.

Eating disorders happen at every body size

One of the most persistent myths in eating disorder care is that you have to look unwell to deserve help. This is wrong. Atypical anorexia — full anorexia psychopathology at a ‘normal’ or higher weight — is increasingly recognised as carrying the same medical and psychological risks as low-weight anorexia. Binge eating disorder, the most common eating disorder, usually presents at average or higher weight. Bulimia nervosa typically presents at normal weight by design. Visible thinness is not the threshold; behaviour and distress are.

Roughly 1 in 4 cases of anorexia and bulimia occur in men or boys, and binge eating disorder is closer to gender-balanced. Men are less likely to be picked up by screens, less likely to seek help, and less likely to be referred when they do. If your score is at or above 20 and you are male, the score is just as relevant.

What modern treatment actually looks like

The image of force-feeding is decades out of date. Modern eating-disorder care is gentle, talk-based, and works with you, not against you.

• Enhanced Cognitive Behavioural Therapy (CBT-E, Fairburn) is the first-line evidence-based treatment for adults with bulimia nervosa, binge eating disorder, and most cases of anorexia at non-acute weight. 20-40 sessions.
• Family-Based Treatment (FBT, Lock & Le Grange) is the first-line for adolescents with anorexia. Parents are trained to refeed at home with clinical support. Recovery rates around 50-60% at 12 months, higher with full course.
• Maudsley Anorexia Nervosa Treatment for Adults (MANTRA) is a NICE-recommended option for adult anorexia, focused on the cognitive and interpersonal style that maintains restriction.
• A clinical dietitian + a mental-health clinician working together is the standard team. A GP or family doctor can route both.

Modern guidelines — NICE NG69 (2017, updated 2020) in the UK, the APA Practice Guideline (4th edition, 2023) in the US — agree on the evidence base. Most people who get treated within the first 3 years of illness do well. Most full recoveries happen within 5-7 years of treatment start.

When to seek urgent medical help

Eating disorders are the deadliest category of mental illness — anorexia nervosa in particular has one of the highest mortality rates of any psychiatric diagnosis. Most deaths are from medical complications and from suicide. The following are medical urgencies, not ‘next month’ situations:

• Not eating or drinking anything for more than 24 hours
• Vomiting multiple times a day
• Heart palpitations, fainting, severe dizziness on standing
• Severe muscle weakness, inability to climb stairs
• Suicidal thoughts or self-harm
• Rapid weight loss (more than 1 kg / 2 lb per week) over several weeks

For mental-health crisis in the US, call or text 988 (Suicide & Crisis Lifeline). In the UK, Samaritans 116 123 (free, 24/7). NEDA Helpline US: 1-800-931-2237 (text ‘NEDA’ to 741741 for crisis text line). Beat UK adult helpline: 0808 801 0677.

EAT-26 vs SCOFF — and what to do if you score differently

SCOFF is a 5-question behavioural screen. EAT-26 is a 26-question attitudes screen. They catch different things:

• EAT-26 catches the restrictive-anorexic attitude cluster (Dieting subscale) and the bulimia-preoccupation cluster very well.
• SCOFF catches behavioural patterns — purging, loss of control, recent weight loss — that EAT-26 misses if the person has not internalised the typical attitudes.

If EAT-26 is below 20 but something still feels off, take SCOFF as a second check. If both are negative but you spend hours a day on food rules, that gap is worth a clinical conversation.

Related tests

• SCOFF eating disorder screen — 5 yes/no items, behavioural focus.
• PHQ-9 depression screen — eating disorders co-occur with depression in roughly half of cases.
• GAD-7 anxiety screen — many people use food to manage anxiety.

Limitations

• Attitudes screen, not behavioural screen. A score of 0 paired with daily binge episodes, purging or extreme restriction is still a clinical issue. The official EAT-26 protocol asks about behaviours separately.
• Restrictive-pattern bias. EAT-26 is most sensitive to restrictive attitudes; it can under-detect binge eating disorder, atypical anorexia, ARFID and orthorexia.
• Adults only. EAT-26 was validated in adults and older adolescents. For under-14s, ChEAT (Children’s Eating Attitudes Test) is the youth-adapted version.
• Cultural calibration. Original validation was in North American populations. The 20 cut-off has been replicated in many countries but cut-off values may shift slightly in non-Western contexts.
• Self-report under-counts. Eating disorders push toward secrecy. Honest answers are the most valuable answers.

Privacy

All calculations run in your browser. We never see, log, or store your individual answers. Only an anonymous event (locale, severity band) is sent to a privacy-respecting analytics service. You can refresh, close, or share this page — your answers stay only on this device.

Licence and source

EAT-26 is freely available for clinical and educational use under the original Garner & colleagues release. Current scoring instructions and the official form are maintained at eat-26.com. Source: Garner DM, Olmsted MP, Bohr Y, Garfinkel PE. The Eating Attitudes Test: psychometric features and clinical correlates. Psychological Medicine 1982;12(4):871-878. PMID 6961471.