Key points
- The Autism-Spectrum Quotient (AQ) is a 50-question self-report measure of autistic traits, published in 2001. The AQ-10 is a 10-question short form.
- Different thresholds come from different studies and purposes: 32 in the original research, 26 in a clinic sample, and 6 or more on the AQ-10 for deciding about referral.
- In a large study of adults referred to a specialist service, AQ scores did not predict who was diagnosed as autistic.
- A high AQ score means you share many traits measured by the scale. It is not a diagnosis, and a low score doesn't rule autism out.
What the AQ is
The Autism-Spectrum Quotient, usually shortened to AQ, is a questionnaire that measures how many autistic traits an adult reports. Simon Baron-Cohen and colleagues at the Autism Research Centre in Cambridge published it in 2001 [1]. It was designed for adults of average or above-average intelligence.
The original paper described it as a way to place a person on a continuum of autistic traits, with the potential for screening "remains to be fully explored" [1]. That phrase is worth remembering, because the AQ is now widely used as if it were an autism test. Its authors were more cautious.
AutismTestQuiz does not currently offer the AQ. This article explains it so you can make sense of a score you got elsewhere.
What it asks about
The full AQ has 50 statements. For each one you choose definitely agree, slightly agree, slightly disagree or definitely disagree. The statements cover five areas, 10 statements each [1]:
- Social skill: how comfortable and skilled you feel in social situations.
- Attention switching: how easily you move between tasks or cope with interruption.
- Attention to detail: how much you notice small details, patterns and numbers.
- Communication: how you follow conversation and understand what people mean.
- Imagination: for example, how you engage with fiction and pretend play.
Each answer in the autistic direction scores 1 point, whether you agree slightly or definitely. Some statements are worded so that agreeing scores the point, and others so that disagreeing does. The total runs from 0 to 50 [1].
Because "slightly" and "definitely" score the same, the AQ counts how many traits you report, not how strongly. Someone who slightly agrees with 30 statements gets the same score as someone who definitely agrees with 30.
The thresholds and where they come from
You will see three different numbers quoted for the AQ. They come from different studies that asked different questions.
32 on the full AQ. In the original study, autistic adults scored 35.8 on average and a randomly selected comparison group scored 16.4 [1]. Eighty per cent of the autistic group scored 32 or more, compared with 2% of the comparison group [1]. The 32 figure describes how those two groups separated. It was not set as a diagnostic cut-off.
26 on the full AQ. A later study gave the AQ to 100 adults referred to a diagnostic clinic. In that group, 26 worked best for flagging people who went on to be diagnosed [2]. The authors called this a preliminary study.
6 on the AQ-10. The AQ-10 is a 10-question short version, built by choosing the questions that best separated a large group of autistic volunteers from non-autistic volunteers. At a score of 6, it identified 88% of autistic adults and correctly excluded 91% of non-autistic adults in that research sample [3]. In England, NICE guidance suggests considering the AQ-10 for adults with possible autism who do not have a moderate or severe learning disability, and offering a comprehensive assessment to people who score 6 or above, or when autism is suspected on clinical judgement [5].
What happened in a real referral clinic
The most useful test of a screening questionnaire is the setting where it will be used. In 2016, researchers looked at 476 adults seen at a national specialist service in the UK for suspected autism [4]. All of them completed the AQ, and all had a full assessment.
Of those adults, 73% were diagnosed as autistic. Their AQ scores did not significantly predict who would be diagnosed [4]. Among people who scored below the AQ threshold, 64% were in fact diagnosed as autistic [4]. The scale also gave many high scores to people who were not diagnosed. The researchers found that generalised anxiety may raise AQ scores and look like autistic traits on a questionnaire [4].
Their conclusion was that the AQ had limited use for deciding who should be assessed in that service, and that guidance recommending it may need to be reconsidered [4].
Why the results differ so much
The first studies compared people who were clearly autistic with people who clearly were not. Under those conditions, the AQ separates the groups well. A referral clinic is different: nearly everyone has come because they have real difficulties, and many have anxiety, depression or other conditions that overlap with autistic traits [4].
There is a second reason. The AQ asks you to judge yourself. Adults who have spent years learning to fit in may not see their own traits clearly, or may describe how they behave now after years of practice rather than what comes naturally. Others may recognise every statement because they are anxious, exhausted or already convinced of the answer.
How to read your own AQ score
- A high score means you reported many of the traits the AQ measures. That is worth taking seriously, especially if those traits have been there since childhood and affect your daily life.
- A score below 26 or 32 doesn't rule autism out. In the clinic study, most people below the threshold were autistic [4].
- Look at the five areas, not just the total. "I score high on attention switching and communication but low on attention to detail" tells you more than one number.
- Think about when each trait started. The AQ doesn't ask, but a clinician will.
- If anxiety is a big part of your life, mention it. It can change how you answer.
What to do next
If your AQ score has made you think more seriously about autism, the next step is a fuller picture, not a second questionnaire. Write down specific examples from your current life and from childhood, and if you can, talk to someone who knew you when you were young.
Our Trait Profile asks 59 questions across seven areas and asks whether each experience was true now, as a child, or both, which the AQ does not. It is an original questionnaire with no diagnostic cut-off, so it describes your answers rather than a probability. If you are considering an assessment, our guide to getting assessed as an adult covers the main routes and how to prepare.
Common questions
Do men and women score differently? In the original comparison group, men scored slightly higher on average than women. Among autistic participants, men's and women's scores did not differ significantly [1].
Should I take the AQ-10 or the full AQ? The AQ-10 was built to help health professionals decide whether to refer someone for assessment [3][5]. The full AQ gives a more detailed picture across five areas. Neither is a diagnosis, and the clinic study above suggests neither should be the only thing that decides whether you seek an assessment [4].
Is a score of 31 "almost autistic"? No. The thresholds are not points on a ladder. A score just under 32 means the same as a score just over it: you reported many of the traits the AQ measures. What matters more is which traits, since when, and how much they affect your life.
Why does my score change when I retake it? Your answers can shift with mood, stress, tiredness and how you read the statements on the day. If your score moves by a few points, that is normal. Look at which answers changed, not just the total.
Do scientists score higher? The original study found that students in the sciences, and especially mathematicians, scored higher on average than students in the humanities and social sciences [1]. That describes groups, not individuals, and it doesn't mean a high score reflects your subject choice rather than autistic traits.
Take the related test
Quick check
Ten questions, about two minutes. A first look before the full profile.
Start questionnaireTrait Profile
59 questions about social communication, routines, interests and sensory experience, and when each one began.
Start questionnaireRelated reading
References
- Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & Clubley, E. (2001). The Autism-Spectrum Quotient (AQ): Evidence from Asperger syndrome/high-functioning autism, males and females, scientists and mathematicians. Journal of Autism and Developmental Disorders, 31(1), 5–17.
- Woodbury-Smith, M. R., Robinson, J., Wheelwright, S., & Baron-Cohen, S. (2005). Screening adults for Asperger syndrome using the AQ: A preliminary study of its diagnostic validity in clinical practice. Journal of Autism and Developmental Disorders, 35(3), 331–335.
- Allison, C., Auyeung, B., & Baron-Cohen, S. (2012). Toward brief "red flags" for autism screening: The Short Autism Spectrum Quotient and the Short Quantitative Checklist in 1,000 cases and 3,000 controls. Journal of the American Academy of Child & Adolescent Psychiatry, 51(2), 202–212.
- Ashwood, K. L., Gillan, N., Horder, J., Hayward, H., Woodhouse, E., McEwen, F. S., et al. (2016). Predicting the diagnosis of autism in adults using the Autism-Spectrum Quotient (AQ) questionnaire. Psychological Medicine, 46(12), 2595–2604.
- National Institute for Health and Care Excellence (2012, updated 2021). Autism spectrum disorder in adults: diagnosis and management (CG142).