How accurate is the AQ-10?

Short answer: the AQ-10 is a well-validated screening tool — good at flagging who should have a fuller evaluation, not at handing out a diagnosis. It was developed by Allison, Auyeung & Baron-Cohen in 2012 and has been tested against clinical interviews in independent studies. A short, NICE-recommended first-pass screen distilled from the longer Autism Spectrum Quotient.

What the numbers mean

Sensitivity is how often the test correctly flags people who do have the condition; specificity is how often it correctly clears those who don't. High figures on both mean the AQ-10 misses few real cases while raising few false alarms — which is exactly what you want from a first-pass screen.

The honest limits

No screening tool is perfect, and self-administration online adds its own caveats:

• A brief 10-item screen misses nuance — many autistic adults score in unexpected ranges.
• It was validated as a referral aid, not a diagnostic instrument.
• Only a specialist assessment can diagnose autism.

Does taking it online change the accuracy?

The questions and scoring here are identical to the clinical version, so the instrument itself is unchanged. What differs is context: a clinician also weighs your history, rules out other causes, and interprets borderline scores. That's why every result here points you toward a professional rather than standing in for one. Reproducing the instrument faithfully is our job; diagnosing is theirs.