The RAADS-R (Ritvo Autism Asperger Diagnostic Scale, Revised) is an 80-question screening tool designed to help identify autism spectrum disorder in adults. A total score of 65 or higher is consistent with an ASD diagnosis, while 64 or lower is not. It was developed as a clinician-assisted questionnaire, meaning you answer the questions yourself, but a trained professional is meant to be in the room to clarify anything confusing. If you’ve seen it floating around the internet as an online quiz, that’s a significant departure from how it was designed to work.
What the Test Measures
The RAADS-R breaks its 80 questions into four areas of functioning that are commonly affected in autism:
- Social relatedness (39 items): Difficulty reading social cues, forming relationships, or understanding unspoken social rules.
- Sensory-motor symptoms (20 items): Unusual responses to sounds, textures, lights, or repetitive physical movements.
- Circumscribed interests (14 items): Intensely focused hobbies or topics that dominate your attention.
- Language (7 items): Differences in how you use or interpret language, such as taking things very literally.
Each question asks you to rate whether a behavior or experience applies to you now, applied when you were younger, or has never applied. This structure is intentional. Many adults learn to mask autistic traits over time, so the test tries to capture patterns from across your life rather than just a snapshot of how you present today.
How Scoring Works
Scores on the RAADS-R can range from 0 to 240. In the international validation study, autistic adults scored between 44 and 227, while non-autistic comparison groups scored between 0 and 65. The threshold of 65 was chosen because it produced the best separation between the two groups: 97% sensitivity (correctly identifying people who do have autism) and 100% specificity (correctly ruling out people who don’t), for an overall accuracy of 98.5%.
A Swedish validation study found slightly different numbers, with 91% sensitivity and 93% specificity at a cutoff of 72. The fact that different studies land on somewhat different thresholds is normal for any screening tool and is one reason the score is meant to be interpreted by a clinician alongside other information, not taken as a standalone verdict.
Who It’s Designed For
The RAADS-R was built specifically for adults with average or above-average intellectual ability who suspect they may be autistic but were never diagnosed as children. This is a large and growing group. Many people, particularly women and those who learned to camouflage social difficulties, reached adulthood without anyone flagging their traits. The test is not validated for children or for adults with significant intellectual disabilities.
It’s also not a general population screener. It was designed for people who already have some reason to wonder about autism, typically as one piece of a broader clinical evaluation. Think of it as a structured way for a clinician to quantify traits you describe, not as a first-pass filter for the general public.
Clinician-Assisted vs. Self-Report
This is where the gap between the test’s design and its real-world use gets important. The RAADS-R was created to be filled out by you while sitting with a trained clinician who can explain questions, probe your answers, and catch misunderstandings. Many of the questions involve subtle social or sensory experiences that are genuinely hard to evaluate about yourself, especially if you’ve never had a framework for understanding them.
In practice, many clinics started mailing the questionnaire to patients to complete at home before their appointment, treating it as a pure self-report measure. One study that examined this approach at an adult autism service found that the RAADS-R “lacks predictive validity” when used as a self-report screening tool without clinician involvement. In other words, it didn’t reliably distinguish between people who would and wouldn’t ultimately receive an autism diagnosis.
Online versions of the test remove the clinician entirely. You can find the RAADS-R on several websites where you answer the 80 questions and get an instant score. While this can be a useful starting point for self-reflection, the accuracy figures (97% sensitivity, 100% specificity) do not apply to that context. Those numbers came from controlled settings with professional guidance.
What a High Score Means
A score above 65 does not mean you are autistic. It means your self-reported traits are consistent with the pattern seen in adults who have been clinically diagnosed. The distinction matters. Several other conditions share traits with autism, including social anxiety, ADHD, PTSD, and certain personality disorders. Difficulty with social situations, sensory sensitivities, and intense focus on specific interests can all show up in those contexts too, which is exactly why the test was meant to be one component of a larger evaluation.
If you took the test online and scored above the threshold, the most useful thing to do with that information is bring it to a clinician who specializes in adult autism assessment. It gives them a structured starting point. If you scored below 65, it’s worth knowing that the test is better at confirming autism than ruling it out in all cases, particularly for people who have spent years learning to suppress or work around their traits.
How It Compares to Other Screening Tools
The RAADS-R is one of several autism screening instruments used for adults. The Autism Quotient (AQ), a 50-question self-report measure, is another common one. Both are often used together in clinical settings. The RAADS-R is longer and more detailed, which can be an advantage for capturing nuance but also makes it more time-consuming and potentially more confusing without guidance.
Neither tool is diagnostic on its own. A full autism assessment typically involves a detailed developmental history (often including information from family members), observation of social communication patterns, and ruling out other explanations for the traits in question. The RAADS-R contributes structured data to that process, but it was never intended to replace it.