RAADS-R Test Accuracy and Reliability: What the Research Actually Shows

Why People Ask About RAADS-R Accuracy

When adults take an online autism screener, a natural question follows: "Can I trust this result?" It is a fair question. A score is only useful if the tool producing it is accurate, and autism is a complex diagnosis that no single questionnaire can deliver on its own.

This article reviews the research on RAADS-R test accuracy and reliability, explains the difference between a screening tool and a diagnostic instrument, and outlines the situations where the test performs well — and where it can mislead.

Screening vs Diagnosis: A Critical Distinction

The most important thing to understand is that the RAADS-R is a screening tool, not a diagnostic instrument.

A screening tool identifies people who are likely to have a condition and would benefit from further evaluation. It casts a relatively wide net.
A diagnostic instrument is used by trained clinicians, alongside interviews and observation, to determine whether someone meets diagnostic criteria.

The RAADS-R was designed to help clinicians decide whether a full autism evaluation was warranted. It was never meant to deliver a diagnosis on its own, no matter how high the score.

Original Validation Research

The RAADS-R was validated in a 2011 study by Ritvo and colleagues published in the Journal of Autism and Developmental Disorders. The study compared:

201 adults with autism spectrum disorder diagnoses.
578 adults with other psychiatric conditions (anxiety, depression, ADHD, eating disorders, and others).
590 adults with no psychiatric diagnosis.

Key findings from the original validation:

Sensitivity: 97 percent. This means 97 percent of the autistic adults scored above the cutoff — the test correctly identified almost all of them.
Specificity: 100 percent. None of the adults without autism scored above the cutoff in this study.
Cutoff score: 65. Scores of 65 or above were considered a positive screen.

These numbers are unusually high, and later research has painted a more nuanced picture.

What Later Studies Found

Subsequent research has produced more realistic estimates of the test's performance, especially in general clinical populations.

Sensitivity Remains High

Most replication studies find that the RAADS-R is good at identifying autistic adults who take it. Sensitivity estimates in real-world clinical samples generally fall between 70 and 95 percent depending on the population and setting.

Specificity Is Lower Than Originally Reported

The original 100 percent specificity has not held up in broader samples. In particular:

Adults with trauma histories, borderline personality disorder, ADHD, social anxiety, or complex PTSD can score above the cutoff without meeting autism criteria.
People who are highly introspective or who have spent years in therapy may relate to many items without being autistic.
The test cannot distinguish autism from other conditions that involve social, sensory, or executive differences.

This means a high score is a strong signal to seek evaluation, not a definitive answer.

Test-Retest Reliability

Test-retest reliability — the degree to which your score is stable if you take the test twice — is generally moderate to good for the total score, but lower for individual domain scores. This is one reason clinicians focus more on the total score than on the breakdown.

Factors That Affect Accuracy

Several factors can skew results:

Masking and Camouflaging

If you minimize or hide your traits, even unconsciously, your score may be lower than your actual trait level. This is especially common in women, people diagnosed late, and people with strong verbal skills.

Co-Occurring Conditions

Anxiety, depression, ADHD, trauma, and personality disorders can inflate scores. A clinician's job is to disentangle which traits come from which condition.

Self-Awareness and Insight

The RAADS-R relies on self-report. People who struggle with introspection or alexithymia may underreport. People who overthink or pathologize normal variation may overreport.

Answer Style

The four response options ("true now and when young," "true now only," "true only when young," "never true") can confuse test-takers. Misreading them can produce inaccurate results.

Translation and Cultural Factors

Translations may shift the meaning of subtle items. Cultural norms around eye contact, social behavior, and emotional expression also affect responses.

What a High Score Means — and Does Not Mean

If your RAADS-R score is 65 or higher, here is what the evidence supports:

It is reasonable to suspect you may have meaningful autistic traits.
A formal evaluation with a qualified clinician is appropriate.
The result should be considered alongside your developmental history, lived experience, and any co-occurring conditions.
The score alone does not equal a diagnosis.

If your score is below 65 but you strongly relate to autistic experiences, the test may have undercounted your traits. Masking, learned coping, or atypical presentation can all produce a false negative. A clinician evaluation is still worth pursuing.

How to Use the RAADS-R Well

To get the most out of the test:

Answer honestly, including traits you usually hide.
Consider your whole life, not just the past month.
Reflect on each domain score, not just the total.
Bring your results to a clinician experienced in adult autism.
Do not use the score to self-diagnose in isolation.

How It Fits Into a Diagnostic Process

A typical adult autism evaluation includes:

A clinical interview focused on developmental history.
Standardized instruments such as the ADOS-2 and the ADI-R.
Self-report measures like the RAADS-R, AQ, or CAT-Q.
Collateral information from parents or partners when possible.
Screening for co-occurring conditions.

The RAADS-R is one data point among several. Its value is in prompting deeper evaluation, not replacing it.

Take the RAADS-R

If you would like to take the RAADS-R as a first step, you can take the test here. Treat the result as the beginning of a conversation — with yourself, with a clinician, and with the community — rather than the final word.