Predictive validity of self-report questionnaires in the assessment of autism spectrum disorders in adults.
RAADS-R and AQ screens miss and mislabel too many adult outpatients to stand alone.
01Research in Context
What this study did
The team tested three adult ASD self-report screens in real outpatient referrals. They used RAADS-R, AQ-28, and AQ-10 with adults sent for autism evaluation.
Each person filled out the forms before seeing the clinician. The researchers then checked if the scores matched the final diagnosis.
What they found
All three tools missed too many autistic clients and flagged too many non-autistic clients. The errors were large enough that the screens could not be trusted alone.
In short, the questionnaires gave too many false negatives and false positives for safe use in routine practice.
How this fits with other research
Smit et al. (2019) saw the same poor accuracy with the SCQ in school-age kids at community mental health clinics. Both studies show that popular checklists fail when they leave research settings.
Cary et al. (2024) later added EEG data and still could not separate mild ASD from other adult outpatients using the same AQ scores. Their null result extends the 2015 warning that self-ratings are not enough.
Hartwell et al. (2025) reviewed 37 ASD screens and found no single tool excels on every key point. Their map supports the 2015 finding by showing that high accuracy claimed in labs rarely holds in clinics.
Why it matters
If you work with adult referrals, do not let a low or high RAADS-R or AQ score decide the diagnosis. Use the scores only as one weak clue, then follow with a full interview and observation. This saves clients from false reassurance or unnecessary labels.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Add a structured interview or ADOS module before telling an adult client they do or do not have ASD, even if their RAADS-R score is extreme.
02At a glance
03Original abstract
While various screening instruments for autism spectrum disorders are widely used in diagnostic assessments, their psychometric properties have not been simultaneously evaluated in the outpatient setting where these instruments are used most. In this study, we tested the Ritvo Autism Asperger Diagnostic Scale-Revised and two short versions of the Autism-Spectrum Quotient, the AQ-28 and AQ-10, in 210 patients referred for autism spectrum disorder assessment and in 63 controls. Of the 210 patients, 139 received an autism spectrum disorder diagnosis and 71 received another psychiatric diagnosis. The positive predictive values indicate that these tests correctly identified autism spectrum disorder patients in almost 80% of the referred cases. However, the negative predictive values suggest that only half of the referred patients without autism spectrum disorder were correctly identified. The sensitivity and specificity of each of these instruments were much lower than the values reported in the literature. In this study, the sensitivity of the Ritvo Autism Asperger Diagnostic Scale-Revised was the highest (73%), and the Autism-Spectrum Quotient short forms had the highest specificity (70% and 72%). Based on the similar area under the curve values, there is no clear preference for any of the three instruments. None of these instruments have sufficient validity to reliably predict a diagnosis of autism spectrum disorder in outpatient settings.
Autism : the international journal of research and practice, 2015 · doi:10.1177/1362361315589869