The EAT-26 asks about dieting behaviour, food preoccupation, perceived pressure to gain weight and control over eating, each item rated on a six-point frequency scale (from "always" to "never") and scored 0-3, giving a total of 0-78. It carries three subscales from the original factor analysis: Dieting; Bulimia & Food Preoccupation; and Oral Control. Alongside the 26 items, the standard form adds behavioural questions (bingeing, purging, laxative use, treatment history and recent weight loss) which flag referral regardless of the total score.
02 - Scoring & cutoffs
How scoring works.
A total of 20 or more is the standard referral threshold: it indicates a level of eating-related concern that warrants assessment by a qualified professional. A positive behavioural question does the same even when the total is below 20. The EAT-26 is a screen, not a diagnostic instrument. It identifies risk, and diagnosis requires clinical interview. Scores respond to symptom change, but the instrument's evidence base is screening, not outcome measurement.
03 - Validation evidence
How well it performs.
The EAT-26 was derived from the 40-item original by factor analysis, correlating 0.98 with the EAT-40 (Garner et al. 1982). Mintz and O'Halloran (2000) validated it against DSM-IV eating disorder criteria in nonclinical women, supporting the cutoff of 20 for discriminating those meeting diagnostic criteria from those not. Known caution: in nonclinical and dieting populations the EAT-26 produces a substantial false-positive rate, so positive screens over-call and always need clinical follow-up.
0.98
Correlation with EAT-40
Garner et al., 1982
>= 20
Referral cutoff
Mintz & O'Halloran, 2000
0-78
Score range
26 items scored 0-3
26
Items
Plus behavioural questions
04 - How it compares
How it compares to the alternatives.
Instrument
Items
Time
When to reach for it
EAT-26
26 + behavioural
~15 min
The reference disordered-eating attitudes screen; copyrighted, so items are not reproduced here. Use it when a validated multi-attitude measure is wanted and licensing is in place.
Free. A five-question referral gate for eating disorders; quick and reproducible with attribution.
05 - When to use it
Right tool, wrong tool.
Reach for it when
-Screening for disordered eating attitudes in adults and older adolescents in primary care, student health and eating-disorder services
-Triaging who warrants a fuller eating-disorder assessment when face-to-face evaluation is limited
-Tracking broad symptom attitudes over treatment where a licensed copy is already in use
-Settings where the behavioural questions (bingeing, purging, laxatives) add a referral flag the total alone misses
Reach for something else when
-Diagnosing an eating disorder; the EAT-26 is a screen, and diagnosis requires clinical interview
-Measuring symptom change precisely; it is screening-validated, not an outcome instrument
-Populations where the high false-positive rate in nonclinical samples would overwhelm follow-up capacity
-Any commercial platform reproducing the items without a licensing agreement from Eating Attitudes, LLC
06 - Confidence & precision
Reading the score with care.
The EAT-26 has no established minimal clinically important difference; the cutoff of 20 is a screening threshold, not a change metric. Internal consistency is high (Cronbach's alpha around 0.90 in the original clinical validation), but the instrument's evidence base is screening rather than outcome measurement, so score changes between administrations should be interpreted cautiously. The three subscales are less well replicated than the total score.
07 - Limitations
What it cannot tell you.
In nonclinical and dieting populations the EAT-26 produces a substantial false-positive rate, so positive screens over-call and always need clinical follow-up.
The total does not distinguish between anorexic and bulimic presentations; the subscales are imperfect and less replicated than the total score.
It was developed and validated largely in young women, so performance in men, older adults and non-Western populations is less established.
The behavioural questions depend on honest self-report of shame-laden behaviours (purging, laxative use) and under-report.
A screen, not a diagnosis: the EAT-26 indicates risk that warrants assessment. It does not confirm an eating disorder.
08 - Licensing, explained
How licensing works.
The EAT-26 is protected by copyright. Its originators waive all fees and royalties for individual users, but reproducing the test on paper or electronically requires a permission letter from the copyright holder, and commercial web platforms must obtain a licensing agreement from Eating Attitudes, LLC. We therefore do not reproduce the items or offer a calculator. The official test, scoring instructions and permission requests are available free at eat-26.com. If you need a freely reproducible alternative for quick screening, the SCOFF is five questions and free to use with attribution.
See how Aisel removes friction where it costs most. A 20-minute walkthrough tailored to your clinic.
We value your privacy
We use cookies to analyse site usage and improve your experience. Analytics and embedded media (e.g. YouTube) only load if you accept. Read our cookie policy.