How accurate is the RAADS-R?

Accuracy for a questionnaire has two sides. Sensitivity is how often autistic people score above the threshold. Specificity is how often non-autistic people score below it. A tool can be excellent at one and poor at the other, and both numbers depend heavily on who was in the study. The RAADS-R is a clear example of how much the setting matters.

What the original study reported

These are remarkable figures, and they explain why the RAADS-R is so often described as highly accurate. But the autistic group had diagnoses already confirmed with structured tools, the comparison group was clearly separate, and a clinician sat with every participant. That design answers the question "can this scale tell confirmed autistic adults apart from others?" It does not answer the harder question most readers have: "can it sort out the adults who walk into a clinic unsure?"

Other validation studies

Translations in other countries have broadly supported the scale's reliability, while finding lower accuracy than the original.

Studies in real referral clinics

The most relevant evidence comes from people referred for assessment, because that is the group most likely to use the scale. Here the picture changes.

Why accuracy drops in clinics

• People referred for assessment usually have real difficulties, whatever the cause, so they endorse many statements.
• Anxiety, ADHD, trauma, low mood and personality differences share features with autism on self-report.
• Some autistic adults, especially those who mask, under-report traits or are unsure about childhood.
• Taking it without a clinician removes the chance to clarify confusing statements.

Is the RAADS-R a good test?

It is a carefully designed questionnaire with strong reliability and a sensible focus on lifelong traits. It is a weak decision tool on its own when the people taking it are already struggling. Many clinicians value it for the conversation it opens: which statements a person marked, what examples they give, and how answers differ between now and childhood.

Is it still used?

Yes. UK NICE guidance lists the RAADS-R as one of several formal tools that may help in complex adult assessments, and many services use it as part of their paperwork. What has changed is how much weight clinicians put on the total. The research above is one reason services increasingly treat it as background information rather than a gate.

Questions to ask about any accuracy claim

Websites often quote the 97% and 100% figures without context. When you see an accuracy number for any autism questionnaire, a few questions help you judge it.

• Who was tested? Confirmed autistic adults versus volunteers is an easier comparison than a mixed clinic waiting list.
• Who was in the comparison group? People with ADHD, social anxiety or trauma are far harder to separate from autistic adults than people with no diagnosis.
• How was it taken? With a clinician present, on paper, or alone online?
• Was the threshold chosen and tested in the same sample? Results usually look better the first time a cut-off is set.

Applied to the RAADS-R, these questions explain most of the gap between the original paper and later clinic studies. None of this makes the scale useless. It makes it a tool whose numbers need a professional to put them in context.

What this means for your score

If you scored 65 or above, your answers resemble those of autistic adults in the original study, and that is reasonable grounds to explore an assessment. It is not proof. If you scored below 65, that does not rule autism out, especially if you have spent years adapting. Either way, a qualified professional is the only person who can diagnose autism, and they will want far more than a number.