Skip to content

About This Autism Test: How It Was Built and What It Can Do

The design decisions behind the 40-question test, what it measures, how it is scored, what it deliberately does not claim, and how it compares with the instruments used in research and clinics.

40 questions6 trait areas5-point scale7 reverse-scored itemsNo account

What this test is

Welcome. This test is a 40-item self-report questionnaire measuring autistic traits in adults. The test produces a score out of 100 and six separate area scores. It takes about eight minutes and asks for just your answers, no email sign up, and no fee.

It is a screening questionnaire. That word has a specific meaning. Screening sorts people into “probably worth a closer look” and “probably not”, and it does so imperfectly in both directions. It is not assessment, and the distance between the two is larger than most people assume.

Where the questions came from

Every item was written from scratch for this site. None is taken, adapted or paraphrased from the AQ, the AQ-10, the RAADS-R, the CAT-Q or any other published instrument.

There is a practical reason and a principled one.

The practical reason is licensing. The AQ family belongs to the Autism Research Centre at Cambridge and is licensed for non-commercial use, with adaptation explicitly prohibited. The RAADS-R is published under a Creative Commons Non-Commercial licence. A great many websites reproduce both anyway. We would rather not be one of them.

The principled reason is that copying an instrument badly is worse than writing a fresh one honestly. Online copies routinely drop reverse-scored items, flatten the response options, or present a threshold as a verdict. A clearly labelled original questionnaire is more useful than a mangled copy of a validated one.

What the items share with the established instruments is structure rather than wording. The trait areas are the ones decades of research keep converging on. Nobody owns those.

How the items are organised

Area Items Maps onto
Social communication 10 Diagnostic criterion A, social communication and interaction
Routine and change 6 Criterion B, insistence on sameness
Sensory experience 7 Criterion B, sensory hyper- or hypo-reactivity
Focused interests 5 Criterion B, restricted and fixated interests
Masking and recovery 6 Not in the criteria; added deliberately
Starting and planning 6 Not in the criteria; commonly co-occurring

The first four map onto the formal diagnostic criteria. The last two do not, and that is intentional.

Why masking is included

Masking is absent from the diagnostic criteria and absent from the AQ and the RAADS-R. It is also the single clearest reason those instruments under-detect women and adults who reached adulthood undiagnosed.

The mechanism is straightforward. Someone who has spent twenty years studying how other people behave socially, and copying it, will answer the social questions as though conversation is manageable, because by now it is rehearsed. The same person may be completely drained by it. A questionnaire that only asks about outcome records the first fact and misses the second.

Separating masking into its own area means those two can move independently, and the gap between them becomes visible rather than being averaged away.

Why executive function is included

Difficulty starting tasks, switching between them, and estimating how long things take is not part of the autism criteria. It is reported so consistently by autistic adults that leaving it out produces an incomplete picture. It also overlaps heavily with ADHD, which is useful information in itself: an elevated score here with little else elevated points at least as strongly toward ADHD as toward autism.

How scoring works

Each answer scores 0 to 4. Forty items gives a maximum of 160, converted to a figure out of 100. Each area is reported separately as a percentage of its own maximum, so areas with different item counts remain comparable.

Reverse-scored items

Seven of the forty are worded so that agreement indicates fewer traits rather than more. There is one in every area.

This guards against acquiescence bias, the well-documented tendency to agree with whatever statement is put in front of you. Without reverse items, someone tapping steadily down the right-hand column would score 100. With them, that pattern produces 82. Reaching the top of the scale requires answering consistently with the trait direction across all forty, including the seven that run the other way.

The practical range

Response pattern Score
Strongly agree with everything 82
Neutral throughout 50
Strongly disagree with everything 18
Answering at random around 50
Fully consistent with the trait direction 100

Most genuine results land between 25 and 75.

What this test does not claim

This section matters more than the rest of the page.

It is not validated. There is no normed sample behind it, no sensitivity figure, no specificity figure, and no published study. The AQ and the RAADS-R have all of those and this does not. If you want a score that can be compared against research literature, use one of those instead.

It has no cut-off. Every band describes how many traits you reported, not whether you are autistic. A threshold would imply detection the instrument cannot deliver. The validated instruments have thresholds because they were tested against thousands of people over years, and even those thresholds perform poorly outside the settings they were developed in.

It cannot separate autism from its lookalikes. Neither can any self-report questionnaire. ADHD, social anxiety, complex trauma, long-term burnout and chronic illness all produce elevated scores.

It cannot rule autism out. A low score is weak evidence, for the masking reason above.

So what is it for? Three things. Giving structure to experiences you have probably noticed for years. Showing you which kinds of experience cluster for you, through the six-area profile. And producing the concrete examples that are genuinely useful if you decide to talk to a clinician. The number itself is the least valuable output.

How it compares

This test AQ AQ-10 RAADS-R
Items 40 50 10 80
Time 8 min 10 to 15 min 2 min 15 to 20 min
Response scale 5-point 4-point 4-point 4-point lifetime
Reverse items Yes, 7 Yes Yes Yes
Subscales reported 6 5 None 4
Asks about masking Yes No No No
Lifetime framing Partial No No Yes
Published validation No Yes Yes Yes
Cut-off None 32 6 65

The honest summary: this test covers more ground and reports a richer profile, and the other three have research behind them that it does not. Which matters more depends on what you want from the result. Our RAADS-R guide goes into how well that instrument actually performs when tested independently, which is relevant context for how much a validation study is worth on its own.

Who it is for

Adults, roughly eighteen and over. The wording assumes working life, long-term patterns and years of accumulated self-knowledge.

It is not suitable for children. For anyone under about thirteen, a parent-report instrument is the right tool, because young children cannot compare their inner experience against anybody else’s. Teenagers sit awkwardly in between, since adolescence produces many of these experiences in people who are not autistic.

If you are a woman who suspects you have been masking for a long time, our female autism test weights those items more heavily.

Privacy

Your answers are stored in your own browser for the length of your visit so the score can be calculated. Nothing is sent to us, there is no account, and we do not ask for your name, your email address or anything else that identifies you. Close the tab and the answers go with it.

One last thing before you start. The most useful thing you can do while answering is notice which statements make you pause. The ones that produce a flicker of recognition are worth writing down, whatever the final number says. Those are the examples a clinician can actually work with.

Forty questions, about eight minutes, instant result.