AQ-10 Scores Explained: A Complete Guide
The AQ-10 is a 10-item adult autism screening developed by Allison, Auyeung & Baron-Cohen (2012). It produces a single total score from 0 to 10, with a clinical cut-off at 6. This guide explains what every score means, where it falls in the population distribution, and what to do next at each level.
Score Bands at a Glance
| 0–3 | Very low to typical low | Below the AQ-10 cut-off. Most adults without autism score in this range. Population mean ~2. |
| 4–5 | Sub-threshold | Some traits endorsed but below the cut-off of 6. Includes the 'broader autism phenotype' range. |
| 6 | At cut-off | Sensitivity 88%, specificity 91% at this score. Threshold for further specialist assessment. |
| 7–8 | Above cut-off | Clearly above the threshold. Strong reason to consider clinical assessment. |
| 9–10 | Very high | Top of the AQ-10 range. Very strong screen-positive. |
What the Research Says
The AQ-10's cut-off of 6 was selected by Allison, Auyeung & Baron-Cohen (2012) from validation data on 1,000 autistic adults and 3,000 controls. At this threshold, sensitivity was 88% (catching 88% of autistic adults) and specificity was 91% (correctly clearing 91% of non-autistic adults). The UK's NICE guidelines adopted this cut-off for primary-care screening.
Adults' AQ-10 scores are remarkably stable on retest — the test-retest reliability is high — so a one-off score is a reasonably faithful representation of what you'd score on retake. Variation across attempts is usually 0–1 points, not enough to move you across the cut-off.
Cross-validation with the longer RAADS-R is the standard practice when an AQ-10 result is borderline. The two screens converge well: an AQ-10 of 6 typically corresponds to RAADS-R 100–130; an AQ-10 of 9–10 typically corresponds to RAADS-R 140+. Disagreement between the two is itself diagnostically interesting.
Per-Score Interpretations
- AQ-10 score 0 interpretation
- AQ-10 score 1 interpretation
- AQ-10 score 2 interpretation
- AQ-10 score 3 interpretation
- AQ-10 score 4 interpretation
- AQ-10 score 5 interpretation
- AQ-10 score 6 interpretation
- AQ-10 score 7 interpretation
- AQ-10 score 8 interpretation
- AQ-10 score 9 interpretation
- AQ-10 score 10 interpretation
Frequently Asked Questions
What's the AQ-10 cut-off score?
6. At this threshold, sensitivity is 88% and specificity is 91% in the original validation.
Is the AQ-10 a diagnosis?
No — only a clinician can diagnose autism. The AQ-10 is a brief screening tool that flags adults for further assessment.
How is the AQ-10 scored?
Items 1, 7, 8, 10 score 1 if you 'definitely' or 'slightly' agree. Items 2, 3, 4, 5, 6, 9 score 1 if you 'definitely' or 'slightly' disagree. Total ranges 0–10.
AQ-10 vs RAADS-R — which?
The AQ-10 is faster (3 min vs 20 min) with strong primary-care evidence. The RAADS-R is more comprehensive — 80 items across four sub-scales. Many clinicians use AQ-10 as initial screen, RAADS-R for depth.
Why might my AQ-10 score change?
It's quite stable on retest. Variation usually reflects how you read ambiguous items more than underlying change. Significant retest differences (3+ points) are uncommon.
Does the AQ-10 work for all ages?
It's validated for adults 16+. For younger people, the AQ-Child or AQ-Adolescent are the appropriate measures.
Are some items more important than others?
All ten items contribute equally — each scores 0 or 1. The total is what's interpreted clinically, not individual items.
Can the AQ-10 be wrong?
Yes — sensitivity 88% means about 1 in 8 autistic adults score below the cut-off; specificity 91% means about 1 in 11 non-autistic adults score above it. It's a good screen, not a perfect one.