No online questionnaire can diagnose an eating disorder on its own, but several validated screening tools do a reasonable job of flagging when something is wrong. The most widely studied is the five-question SCOFF, which catches roughly 84 to 86 percent of eating disorder cases in adults when two or more answers are positive. Screening is just the first filter, though. What comes after it, and what the tests miss, matters at least as much as the score itself.
What the Main Screening Tests Actually Ask
Most “eating disorder tests” you find online are adaptations of a handful of questionnaires developed for clinical or research use. The three you are most likely to encounter are the SCOFF, the EAT-26, and the EDE-Q.
The SCOFF is the shortest. Its name is a mnemonic for five yes-or-no questions about whether you make yourself Sick (vomit), worry you’ve lost Control over eating, have recently lost One stone (about 14 pounds) in three months, believe you’re Fat when others say you’re thin, and whether Food dominates your life. A score of two or more “yes” answers is the standard cutoff for concern. A systematic review and meta-analysis across 25 studies found the SCOFF had a pooled sensitivity of 86 percent and specificity of 83 percent at that threshold, meaning it correctly identifies most people with an eating disorder while also correctly clearing most people without one.1PubMed Central. Eating Disorder Screening: a Systematic Review and Meta-analysis of Diagnostic Test Characteristics of the SCOFF A separate evidence review for the US Preventive Services Task Force reported similar numbers in adults: pooled sensitivity around 84 percent and specificity around 80 percent.2JAMA. Screening for Eating Disorders in Adolescents and Adults: Evidence Report and Systematic Review for the US Preventive Services Task Force
The EAT-26 (Eating Attitudes Test) is longer, with 26 items scored on a scale from “always” to “never.” It covers dieting behavior, food preoccupation, and oral control. A score of 20 or higher is generally considered a flag. A meta-analysis of its reliability found a pooled internal consistency (Cronbach’s alpha) of 0.85, which suggests the questions hang together well as a group.3PubMed Central. A Reliability Generalization Meta-Analysis of the Eating Attitudes Test 26 (EAT-26) Scale That said, the EAT-26 has known weak spots. A Rasch analysis found that seven of its items didn’t fit the measurement model well and that the test is biased toward catching people at moderate to high risk while being less sensitive at the lower end.4PubMed Central. Psychometric properties of the 26-item eating attitudes test (EAT-26): an application of rasch analysis So if you score low, it does not necessarily mean you are fine; it means the test was not designed to detect subtler problems.
The EDE-Q (Eating Disorder Examination Questionnaire) is more detailed and is often used in treatment settings. It measures four dimensions: dietary restraint, eating concern, shape concern, and weight concern. A systematic review found reasonable evidence that EDE-Q scores can distinguish between people with and without eating disorders, though the research base for that comparison relied on fairly small clinical samples.5PubMed Central. Psychometric Evaluation of the Eating Disorder Examination and Eating Disorder Examination-Questionnaire: A Systematic Review of the Literature
Where Screening Tests Fall Short
These questionnaires are screening tools, not diagnostic instruments. The distinction matters. A screen is designed to cast a wide net and sort people into “probably worth investigating” and “probably okay.” It is not designed to tell you which eating disorder you have, or whether you have one at all, with clinical certainty.
Even the SCOFF, which performs well on average, showed much lower sensitivity in at least one study of a multiethnic general population sample, dropping to about 54 percent at the same cutoff that normally catches most cases.6PubMed Central. Validation of the SCOFF Questionnaire for Eating Disorders in a Multiethnic General Population Sample That study also found the positive predictive value was only about 41 percent, meaning that among people the SCOFF flagged, fewer than half actually had a diagnosable eating disorder. The flip side was encouraging: its negative predictive value was 96 percent, so a low score was a fairly reliable all-clear. But the takeaway is that a “positive” result on a screening test is the beginning of a conversation, not a diagnosis.
Screening tools also vary in how well they detect specific disorders. One study testing a detailed screen in college-age women found sensitivity ranged from 90 percent for anorexia nervosa all the way down to 55 percent for purging disorder.7PubMed Central. A Screening Tool for Detecting Eating Disorder Risk and Diagnostic Symptoms among College-Age Women Some newer digital screening tools are trying to close these gaps. One called the IOI-S demonstrated strong test-retest reliability and distinguished probable eating disorders with about 83 percent sensitivity and 90 percent specificity, while also detecting two levels of risk below the diagnostic threshold.8PubMed Central. Identification of high risk and early stage eating disorders: first validation of a digital screening tool That stepwise approach, flagging emerging problems before they become full-blown disorders, may be more useful than a binary “at risk” or “not at risk” result.
There is also an argument that the field needs to improve existing tools rather than keep inventing new ones. Some researchers have pushed for a sharper focus on boosting the sensitivity of current screens, so fewer people slip through, alongside better training for the clinicians who act on the results.9PubMed. Beyond screening in primary practice settings: Time to stop fiddling while Rome is burning
Behavioral Patterns Worth Paying Attention To
If you are reading this article, you probably already have a sense that something is off. Screening questionnaires formalize that sense, but the patterns they are probing for are recognizable without a score sheet. Here are the kinds of changes that tend to show up:
- Rigidity around food: Rules about what, when, or how much you eat that feel impossible to break, even when circumstances change. This goes beyond healthy preferences. It is the distress or panic when the rules cannot be followed.
- Compensatory behavior: Exercising specifically to “make up” for eating, using laxatives or diuretics, or vomiting after meals. Even occasional compensatory behavior is clinically relevant.
- Loss of control: Eating past the point of fullness in a way that feels involuntary, followed by shame or distress.
- Body checking or avoidance: Frequent mirror checking, measuring body parts, pinching skin, or the opposite, going to lengths to avoid seeing your body.
- Social withdrawal around meals: Making excuses to skip meals with other people, hiding food, or eating in secret.
- Cognitive preoccupation: Thinking about food, weight, or body shape to the point that it crowds out other interests and makes it hard to concentrate.
The physical side can be just as telling. Eating disorders affect virtually every organ system, with both immediate and long-term consequences.10PubMed Central. Medical Considerations and Consequences of Eating Disorders Dizziness on standing, hair thinning, feeling cold all the time, loss of menstrual periods, brittle nails, and dental erosion (from repeated vomiting) are among the more common signs. What makes these tricky is that routine blood work often comes back normal even when something is wrong. A study of teenage girls with eating disorders found that biochemical measurements were largely within reference ranges and didn’t reliably track the degree of weight loss or starvation.11PubMed. The significance of routine laboratory analyses in the assessment of teenage girls with eating disorders and weight loss Normal labs do not rule out an eating disorder.
You Don’t Have to Be Underweight
This is one of the most persistent misconceptions and one of the biggest barriers to people recognizing their own problem. Atypical anorexia nervosa involves the same restrictive eating, the same fear of weight gain, and the same medical dangers as “typical” anorexia nervosa, but the person’s weight remains in or above the normal range. Research has found that patients with atypical anorexia can have medical and psychological complications that are similar to, and sometimes more severe than, those seen in patients with the typical form.12PubMed Central. Medical complications and management of atypical anorexia nervosa
A systematic review of adolescents with eating disorders confirmed that medical instability requiring hospitalization occurred in roughly 29 to 42 percent of participants who were not underweight. Rapid weight loss turned out to be a more important marker of medical risk than absolute weight, and there were no significant associations between the degree of underweight and heart rate or temperature.13PubMed Central. Medical instability in typical and atypical adolescent anorexia nervosa: a systematic review and meta-analysis The speed of the drop matters more than where you land on the scale.
The diagnostic system itself has shifted to reflect this. When the DSM-5 replaced the DSM-IV, it loosened some rigid criteria that had been locking people out of a diagnosis. The old “Eating Disorder Not Otherwise Specified” (EDNOS) category, which had been a catch-all for everyone who didn’t quite fit the narrow definitions of anorexia or bulimia, shrank from about 48 percent of cases to 39 percent, while diagnoses of anorexia and bulimia both increased.14Wiley Online Library / International Journal of Eating Disorders. The Impact of DSM-5 on Eating Disorder Diagnoses More people got the specific diagnosis they needed.
Conditions That Don’t Look Like the Stereotype
Eating disorders extend well beyond anorexia and bulimia, and some of the less recognized forms rarely show up on standard screeners.
Orthorexia nervosa is an obsession with “clean” or “healthy” eating that becomes restrictive and distressing. It is not yet a formal diagnosis in the DSM-5, but research suggests it captures something real and distinct. A study using the Eating Habits Questionnaire found that the problematic dimension of orthorexia involved disordered eating symptoms separate from those of other eating disorders, and separate from ordinary healthy eating habits.15PubMed Central. Disentangling orthorexia nervosa from healthy eating and other eating disorder symptoms: Relationships with clinical impairment, comorbidity, and self-reported food choices Interestingly, research distinguishes between “healthy orthorexia” (a genuine interest in nutrition without impairment) and pathological orthorexia nervosa, which was paradoxically associated with relatively unhealthy eating and other lifestyle behaviors despite the person’s stated fixation on health.16PubMed Central. Orthorexia nervosa vs. healthy orthorexia: relationships with disordered eating, eating behavior, and healthy lifestyle choices
Muscle dysmorphia, sometimes informally called “bigorexia,” involves a preoccupation with not being muscular enough. It overlaps heavily with eating disorder symptoms, particularly in men. A systematic review and meta-analysis found that clinicians evaluating someone for muscle dysmorphia should investigate the possible presence of an eating disorder, and vice versa.17PubMed Central. The association between muscle dysmorphia and eating disorder symptomatology: A systematic review and meta-analysis In male bodybuilders specifically, eating disorder symptoms correlated positively with both fat dissatisfaction and muscle dissatisfaction.18PubMed Central. Is There Any Relationship Between Body Image Perception, Eating Disorders, and Muscle Dysmorphic Disorders in Male Bodybuilders? A university study found the prevalence of muscle dysmorphia in male students was about 1.3 percent, comparable to the 1.4 percent prevalence of eating disorders in the same sample, and the two conditions looked similar enough on body image measures that researchers suggested they belong to the same spectrum.19PubMed. Going deeper into eating and body image pathology in males: Prevalence of muscle dysmorphia and eating disorders in a university representative sample
If you spend a lot of time worrying about being too small, structuring your diet obsessively around protein and macros, and feeling genuine distress when you miss a gym session, those patterns deserve the same scrutiny as more traditional eating disorder behaviors. Standard screening tests are not built to detect this.
Risk Factors That Raise the Odds
Eating disorders are not caused by any single personality trait or life event, but certain factors appear repeatedly in the research.
Perfectionism is one of the strongest and most consistent risk factors. A study comparing people at different stages of recovery found that those with active eating disorders and those only partially recovered scored significantly higher on perfectionism than fully recovered individuals and healthy controls. The fully recovered group was indistinguishable from people who never had an eating disorder.20PubMed Central. Perfectionism Across Stages of Recovery from Eating Disorders A separate study comparing eating disorders with depression and OCD found that “concern over mistakes” was elevated across all three conditions, while “personal standards” perfectionism was elevated only in the eating disorder group.21PubMed. Perfectionism in depression, obsessive-compulsive disorder and eating disorders This suggests that eating disorders involve a particular flavor of perfectionism centered on self-imposed standards, not just anxiety about failing.
Trauma and PTSD also co-occur with eating disorders at elevated rates. A network analysis of patients in residential treatment found that PTSD hyperarousal symptoms and negative mood alterations were among the most influential “bridge” symptoms connecting PTSD and eating disorder pathology, suggesting they may help maintain both conditions simultaneously.22PubMed Central. A network analysis of eating disorder, PTSD, major depression, state-trait anxiety, and quality of life measures in eating disorder patients treated in residential care
Autism and ADHD are increasingly recognized as conditions that raise eating disorder risk. Both are associated with health inequalities that include higher rates of disordered eating.23The Lancet Psychiatry. Research priorities for neurodivergent people who experience disordered eating: a Personal View Adults with autism scored substantially higher on measures of eating problems than comparison groups, with mean scores roughly double those of non-neurodivergent adults.24PubMed Central. Food selectivity and eating difficulties in adults with autism and/or ADHD Sensory sensitivities, rigid eating patterns, and difficulties with interoception (the ability to read your body’s hunger and fullness signals) can all play into this overlap, and standard screening tools do not account for it.
Eating Disorders Across Demographics
Another damaging misconception is that eating disorders primarily affect young, thin, white women. A study tracking over 1,100 young women of different ethnic backgrounds found no significant differences in eating disorder prevalence: about 20 percent across white, Hispanic American, African American, and Asian American participants met criteria for a threshold or subthreshold disorder.25PubMed Central. Ethnic differences in eating disorder prevalence, risk factors, and predictive effects of risk factors among young women The study also found no difference in new eating disorder onset across ethnic groups over time, and no evidence that risk factors predicted eating disorders differently by ethnicity.
Gender adds another wrinkle. Research on how the DSM-5 severity system works in practice found that it may not capture severity well for everyone. Among women with bulimia, higher severity categories tracked with increasing eating disorder symptoms as expected. Among men with bulimia, severity categories showed no such association. And for both men and women with anorexia, severity categories based on BMI did not correlate with the actual intensity of eating disorder thoughts and behaviors.26PubMed Central. Gender differences in eating disorder psychopathology across DSM-5 severity categories of anorexia nervosa and bulimia nervosa This means that even the formal diagnostic system, not just the screening tools, can underestimate how sick someone is.
Screening for Younger People
If you are a parent or caregiver wondering about a child or adolescent, the picture looks somewhat different. Young people may not have the self-awareness or vocabulary to report their own symptoms accurately. A study analyzing a large youth mental health survey found that parental report most accurately predicted an eating disorder, correctly identifying cases about 94 percent of the time.27PubMed Central. How does the increase in eating difficulties according to the Development and Well-Being Assessment screening items relate to the population prevalence of eating disorders?
A parent version of the EDE-Q, called the PEDE-Q, was developed to capture the caregiver’s perspective. It showed good psychometric properties and, for restrictive eating disorders like anorexia nervosa, using the higher of the parent and youth scores provided the most accurate diagnostic picture. When diagnosis was based on parent report, significantly more cases of anorexia were identified than when it relied on the young person’s self-report alone.28PubMed Central. Psychometric Properties of the Parent Eating Disorder Examination Questionnaire This is not because adolescents are dishonest. Denial and minimization are hallmarks of restrictive eating disorders at any age, and someone in the grip of one often genuinely does not perceive their eating as disordered.
For avoidant/restrictive food intake disorder (ARFID), which involves extremely limited eating without the body image distortion seen in anorexia, there are newer screening questionnaires for both self-report and parent report. The parent version showed a higher positive predictive value for ARFID diagnosis than the self-report version, again underscoring the value of an outside perspective.29PubMed Central. Validating Online Parent- and Self-Report Screening Methods for Avoidant/Restrictive Food Intake Disorder
What Comes After a Screen
A screening test, whether you take it online or in a doctor’s office, is the beginning of the process. It is not a substitute for professional evaluation. A clinical diagnosis of an eating disorder typically involves a structured interview, a medical examination, and sometimes blood work and other tests, not to confirm the eating disorder (since labs often look normal, as mentioned earlier) but to check for complications like electrolyte abnormalities and heart rhythm changes.30PubMed Central. Detection, evaluation, and treatment of eating disorders the role of the primary care physician
Speed matters here. Early intervention is associated with better outcomes for eating disorders, but help-seeking rates are low and delays before treatment begins are often long.31PubMed Central. A Narrative Review of Early Intervention for Eating Disorders: Barriers and Facilitators Part of the delay is structural, with limited specialist availability and insurance barriers, but part of it is the person not recognizing they need help in the first place. If a screener flags risk, or if you recognize yourself in the behavioral patterns above, that is worth acting on sooner rather than later. Treatment typically involves a team including a physician, a therapist specializing in eating disorders, and a dietitian. The specific approach depends on the diagnosis, severity, and individual circumstances.
The EAT-26 and Special Populations
One underappreciated limitation of popular screening tools is that they were developed and validated on specific populations, and they don’t always work the same way in others. The EAT-26, for example, showed poor psychometric properties when tested in vegetarians and vegans. The test’s items about food restriction and preoccupation can conflate dietary choices made for ethical or environmental reasons with eating disorder symptoms, leading to inaccurate risk scores. The test-retest reliability was also poor in that group, meaning the same person taking it on two occasions could get very different results.32PubMed Central. The 26-Item Eating Attitudes Test (EAT-26): Psychometric Properties and Factor Structure in Vegetarians and Vegans
This doesn’t mean vegetarians and vegans can’t have eating disorders. They certainly can, and some research suggests disordered eating may be more common in these groups. But it does mean that an online EAT-26 result is particularly unreliable if you follow a restrictive diet by choice, and a professional evaluation becomes even more important for getting an accurate picture. The same caution applies to athletes, people in recovery from obesity, and anyone whose relationship with food is shaped by cultural practices that standard Western-developed questionnaires were not built to assess.