AQ-50 test: the Autism Spectrum Quotient screening

The AQ-50 (Autism-Spectrum Quotient) is one of the most widely used autism screening tools in the world. Developed by Professor Simon Baron-Cohen at the University of Cambridge, it measures autistic traits across five key areas: social skills, communication, imagination, attention to detail, and attention switching.

If you have five minutes and want the fastest possible read on where you stand, the AQ-10 is a better starting point than this page. If you have closer to ten minutes and want a domain-by-domain picture rather than one number, stay here and take the full AQ-50 below.

You can take the full 50-question screening right here for free. We also offer the quick AQ-10 (10 questions, about 2 minutes) and the RAADS-R (80 questions designed specifically for adults). All screenings are instant and private.

What the AQ-50 Is

The Autism-Spectrum Quotient was published in 2001 by Simon Baron-Cohen, Sally Wheelwright, Richard Skinner, Joanne Martin, and Emma Clubley in the Journal of Autism and Developmental Disorders. It consists of 50 statements answered on an agree or disagree scale, organized into five domains of 10 questions each: Social Skills, Attention Switching, Attention to Detail, Communication, and Imagination. The AQ-50 was designed to measure the degree to which an adult of average intelligence has autistic traits, whether or not they have a diagnosis, and it became the basis for the shorter AQ-10 a decade later. The AQ-50 is longer than the AQ-10 by design; it was built first, as the comprehensive version, and the shorter form was extracted from it afterward specifically to sacrifice depth for speed. Taking the AQ-50 means answering about how you experience each of the five domains directly, rather than relying on the handful of items the shorter form uses as stand-ins for each area.

What the Five Domains Measure

• Social Skills: Preference for solitary activities, difficulty reading social situations, and comfort or discomfort in group settings.
• Attention Switching: Ease of transitioning between tasks, preference for routine, and tolerance for interruptions.
• Attention to Detail: Noticing patterns, numbers, or small details that others miss, often described as a strength.
• Communication: Comfort with small talk, literal interpretation of language, and knowing when it is your turn to speak.
• Imagination: Preference for factual over fictional content, history with pretend play in childhood, and focused interest patterns.

What each subscale measures

Social Skills asks about preference for solitary activity versus company, comfort in group settings, and how easily you read a social situation while it is happening. A high score here usually means questions and statements about groups, parties, and making new friends felt more true than false.

Attention Switching asks about how easily you move between tasks, how you respond to interruptions, and how much you rely on routine to feel settled. A high score here often goes along with strong distress when plans change without warning and a preference for doing familiar things the same way each time.

Attention to Detail asks about noticing small patterns, numbers, or details that other people tend to miss. Unlike the other four domains, a high score here is often described by autistic and non-autistic people alike as a strength rather than a difficulty, even though it is scored in the same autistic-trait direction as the rest of the questionnaire.

Communication asks about comfort with small talk, how literally you tend to interpret what other people say, and whether you can tell when a listener has lost interest. A high score here often accompanies a preference for direct, literal language over implication or subtext.

Imagination asks about pretend play in childhood, interest in fiction, and how easily you can picture a scenario that has not happened. A high score here does not mean a lack of creativity in general; it specifically reflects the questionnaire's own narrow definition of imaginative flexibility, which some autistic people find does not match how they actually experience their own inner imaginative life.

What if I mask?

Masking can flatten domain scores that would otherwise stand out. A question like "I find social situations easy" is scored by how you answer, not by how much effort that ease costs you, so someone who has trained themselves to appear at ease in groups may score lower on Social Skills than their internal experience would suggest. The same applies to Communication and Attention Switching, where years of practice can make a genuinely effortful skill look automatic on paper. If you recognize this in yourself, it can help to read each of the five domain descriptions above and ask not "can I do this" but "does this take noticeable, ongoing effort for me specifically," since that distinction is usually where masking hides.

How We Score It

Each of the 50 questions is answered on the same four-point scale used on our AQ-10 page: Definitely Agree, Slightly Agree, Slightly Disagree, and Definitely Disagree. For each question, a response in the autistic-trait direction earns a point (either agreeing or disagreeing, depending on how that specific question is scored), and every other response earns nothing. Your total score is a whole number between 0 and 50, split into five domain scores worth up to 10 points each.

Our tool marks a total score of 26 or above as meeting our threshold, and treats a score of 32 or above as an additional, higher band. This is worth stating plainly because it differs from the original research: the 2001 paper by Baron-Cohen and colleagues used 32 or above as its single indicator of clinically significant traits, not 26. We set 26 as an earlier flag on our own scale, above the original study's 32, so that people whose scores fall in that middle range are told their results are worth a closer look rather than being sorted into a simple pass or fail against the published threshold. If you want to compare your result strictly against the original paper, 32 is the number to look at.

How accurate is the AQ-50?

The AQ-50 was designed to test whether autistic traits exist on a continuum across the general population, not to hand any one person a precise accuracy figure. Baron-Cohen and colleagues (2001) built it by comparing groups with and without an existing diagnosis of Asperger syndrome or high-functioning autism, and that comparison is where the study's own 32-point threshold comes from. We do not publish sensitivity or specificity percentages for our implementation, since those numbers describe how a specific research sample performed, not how reliably any single reader's 50 answers describe them.

What the AQ-50 is well suited to is breadth: 50 items across five domains give a fuller picture than the 10-item AQ-10, without requiring the 80-question commitment of the RAADS-R. That breadth is the instrument's real strength, and it is a different kind of value than a claim of clinical accuracy would be.

What the Original Study Found

Baron-Cohen and colleagues developed the AQ-50 to test whether autistic traits exist on a continuum in the general population rather than being present or absent. Their 2001 paper set 32 or above as the threshold indicating a clinically significant level of traits, drawing on comparisons between adults with a diagnosis of Asperger syndrome or high-functioning autism and adults without one. A later paper, by Hoekstra and colleagues, used the same item pool to build the shortened AQ-10.

The five-domain structure was central to the original design. Rather than producing a single undifferentiated score, the authors built the questionnaire so that social skills, communication, imagination, attention to detail, and attention switching could each be examined on their own. This let researchers compare, for example, whether attention to detail (often a relative strength) behaved differently from social skills (often a relative difficulty) across their study population, rather than collapsing both into one number that would obscure the difference.

What It Cannot Tell You

• It is self-report, so it depends on your own insight into your patterns, which can be limited for people who have spent years masking those patterns.
• It is not a diagnosis at any score. A high total does not confirm autism, and a low total does not rule it out.
• It does not gather developmental history. A clinical evaluation looks at how traits showed up in childhood and how much they affect functioning now; a questionnaire answered in one sitting cannot do that.
• It does not distinguish autism from overlapping conditions. Anxiety, ADHD, and OCD can each push scores up on some of these 50 items without the person being autistic.
• Domain scores are a snapshot, not a profile. A low score in one domain and a high score in another describes today's answers, not necessarily a stable pattern across contexts and time.
• It cannot account for how much support you actually need in daily life. Two people can land on the same total score while one manages comfortably and the other struggles significantly; a formal evaluation weighs functional impact, not just trait presence, and that distinction matters for decisions about accommodations or support.

The five domain scores are useful for orientation, not for ranking yourself against anyone else. Two people with the same overall total can arrive at it through very different combinations of domain scores, and neither combination is more or less "autistic" than the other; the domains describe where your traits currently cluster, not how severe your experience is compared with someone else's.

Who the AQ-50 Is For

The AQ-50 suits adults who want more depth than the AQ-10 offers but are not necessarily ready for the 80-question RAADS-R. Because it reports a score for each of the five domains, it is a good fit for people who want to understand which specific areas, such as attention switching or communication, feel most different for them, rather than a single number. It was validated on adults of average or above-average intelligence and is not intended for screening young children; for toddlers, see the M-CHAT-R-based screening for kids.

It is also a reasonable middle step for people who took the AQ-10, scored close to its threshold, and want a second opinion from a longer instrument before deciding what to do next. Because the AQ-10's ten items are drawn from the AQ-50, taking both is not redundant in the way it might sound. The AQ-50 asks about the same broad territory in far more detail, so a domain that felt ambiguous in ten questions often becomes clearer once you have answered ten questions about that domain alone. Students, professionals preparing to request workplace accommodations, and adults comparing their own experience against a partner's or family member's have all found the domain breakdown useful for grounding a conversation in specifics rather than generalities.

What a "Meets" or "Below" Result Means

A total of 26 or above meets our threshold, and a total of 32 or above puts you in the higher band that matches the original study's own cutoff. Either result means your answers describe a level of autistic traits that is worth exploring further, whether through the longer RAADS-R, reading about the domains that stood out for you, or a conversation with a professional. A total below 26 means your answers did not reach either threshold on this occasion, which does not rule autism out on its own, particularly if some of your domain scores were high even while your overall total stayed low, or if you have spent years learning to answer questionnaires the way you think you are expected to rather than the way you actually feel. See how we explain results for more on interpreting domain-level patterns, and our method for the full detail on how every screening on this site is scored.

Frequently Asked Questions