I’m currently undergoing evaluation for ADHD & ASD, and as a part of that, one of the tests I was asked to take was the RAADS-R, and on the page with the test it explains the test’s reliability and validity:

• Sensitivity 97% Test’s ability to identify positive results
• Specificity 100% Test’s ability to identify negative results
• Concurrent validity 96% Test’s validity compared with ADOS Module 4, SRS
• Test–retest reliability .987 Test’s agreement between results of successive measurements

A sensitivity of 97% means that 97% of autistic people who took the test met the cutoff score and were accurately classified as autistic. Said another way, it represents the proportion of autistics who were correctly classified as autistic based on the questionnaire.

In contrast, a 100% specificity score means that all (100% of) the neurotypical people who took the test were below the cutoff score and were accurately classified as non-autistic. It represents the proportion of neurotypicals who were correctly classified as neurotypical based on the questionnaire.

The takeaway is that the RAADS–R has a high probability of accurately classifying those who take the test as autistic or not autistic.

It then cites this article: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3134766/

I’m wondering how suspicious I should feel here by these numbers - I notice in the cited study, re-testing in a UCLA study was with only 15 ASD subjects and another 15 control subjects … Of the studies listed, only one had more than 50 participants, most were fewer than 20 total.

I guess I wonder how confident we should feel that no neurotypical will score above a certain threshold (a score of 65)?