RAADS-r Score Data: 100,968 Real Test Results
Medically reviewed by Dr Neil Drew
Chartered Clinical Psychologist (BSc Hons, MSc, DClinPsy, CPsychol)
Follow us
Follow us
Medically reviewed by Dr Neil Drew
Chartered Clinical Psychologist (BSc Hons, MSc, DClinPsy, CPsychol)
Published by NeuroDirect | Dataset period: August 2025 – July 2026 | n = 100,968
Quick Links
The mean, median (118), and mode (136) all sit well above the RAADS-R clinical threshold of 65 — a finding consistent across a dataset of over 100,000 completions, representing one of the largest community-based RAADS-R samples ever published.
The RAADS-R (Ritvo Autism Asperger Diagnostic Scale–Revised) is an 80-question self-report instrument developed to help clinicians identify autism in adults who may have been missed during childhood. It was validated in a landmark 2011 study by Ritvo et al. across multiple clinical sites in the United States, Sweden, Australia, and the Netherlands, and has since become one of the most widely used structured autism screening tools for adults worldwide.
The test measures autistic traits across four domains:
Scores range from 0 to 240. The validated clinical threshold is 65. The original validation study reported sensitivity of 97% and specificity of 100% in distinguishing autistic from non-autistic adults in a clinical sample. The RAADS-R is a screening tool, not a diagnostic instrument — formal diagnosis requires assessment by a qualified professional. You can take the RAADS-R test here. Another popular option is the Autism Quotient (AQ) Test.
All data is drawn from completed RAADS-R test sessions administered via NeuroDirect.co.uk, a UK-based neurodivergence screening platform. The test is freely accessible to any visitor without registration.
Internal test completions by platform operators were excluded from all analysis.
The RAADS-R on NeuroDirect is self-administered: respondents complete the test without clinician oversight, in their own time, on any device. No demographic information is collected at the point of test completion. The dataset reflects a global, self-selecting population of adults who chose to take the test of their own volition.
n = 100,968 completed tests. The dataset has grown from 72,111 completions to 100,968, with the mean score remaining stable (111.51 → 114.09), suggesting consistent population behaviour rather than compositional drift.
Descriptive statistics were computed in MySQL/MariaDB using native aggregate functions. The median was computed by ordered offset at position FLOOR(n/2). Quartiles were computed using NTILE(4) window function. The mode represents the single most frequently occurring score value.
| Statistic | Value |
|---|---|
| n | 100,968 |
| Mean | 114.09 |
| Median | 118.00 |
| Mode | 136 (800 occurrences) |
| Standard Deviation | 49.37 |
| Variance | 2,437.01 |
| Minimum | 1 |
| Maximum | 240 |
| Range | 239 |
The positive skew between mean (114.09) and mode (136) reflects the right-leaning tail of the distribution. The median of 118 indicates that more than half of all completions scored above the point typically associated with clearly present autistic traits.
| Group | n (Ritvo 2011) | Mean (Ritvo 2011) | NeuroDirect Mean |
|---|---|---|---|
| Autistic adults | 197 | 164.34 | — |
| Non-autistic controls | 233 | 64.59 | — |
| NeuroDirect community | 100,968 | — | 114.09 |
The NeuroDirect community mean of 114.09 sits almost exactly between the autistic (164.34) and non-autistic (64.59) means from the Ritvo validation study — consistent with a self-selecting population of people who have reason to suspect autism but have not yet received a formal diagnosis.
| Quartile | Score |
|---|---|
| Q1 (25th percentile) | 77 |
| Q2 / Median (50th percentile) | 118 |
| Q3 (75th percentile) | 152 |
| Interquartile Range (IQR) | 75 |
The IQR spans 77–152 — both values well above the clinical threshold of 65. Even the bottom quarter of the distribution sits above the clinical threshold.
| Percentile | Score |
|---|---|
| 5th | 26 |
| 10th | 41 |
| 25th | 75 |
| 50th (median) | 116 |
| 75th | 151 |
| 90th | 175 |
| 95th | 187 |
| 99th | 203 |
| Rank | Score | Frequency |
|---|---|---|
| 1 | 136 | 800 |
| 2 | 127 | 797 |
| 3 | 126 | 795 |
| 4 | 141 | 794 |
| 5 | 135 | 786 |
| 6 | 128 | 778 |
| 7 | 129 | 771 |
| 8 | 142 | 764 |
| 9 | 130 | 761 |
| 10 | 147 | 759 |
All ten most common scores fall between 126 and 147 — sitting firmly in The Clear Signal and Deep Profile bands.
n = 100,968 · NeuroDirect.co.uk · Clinical threshold: 65
Score distribution highlights across 100,968 completions:
Full chart visible on desktop · n = 100,968
| Score Band | Total Completions | % of Sample |
|---|---|---|
| 0–9 | 632 | 0.60% |
| 10–19 | 2,336 | 2.23% |
| 20–29 | 4,162 | 3.97% |
| 30–39 | 5,265 | 5.02% |
| 40–49 | 4,001 | 3.82% |
| 50–59 | 4,367 | 4.17% |
| 60–69 | 4,620 | 4.41% |
| 70–79 | 4,890 | 4.67% |
| 80–89 | 5,576 | 5.32% |
| 90–99 | 6,243 | 5.96% |
| 100–109 | 6,536 | 6.24% |
| 110–119 | 6,975 | 6.66% |
| 120–129 ★ | 7,504 | 7.16% |
| 130–139 ★ Peak | 7,530 | 7.19% |
| 140–149 ★ | 7,187 | 6.86% |
| 150–159 | 6,751 | 6.44% |
| 160–169 | 5,987 | 5.71% |
| 170–179 | 5,034 | 4.80% |
| 180–189 | 3,955 | 3.77% |
| 190–199 | 2,601 | 2.48% |
| 200–209 | 1,503 | 1.43% |
| 210–219 | 735 | 0.70% |
| 220–229 | 302 | 0.29% |
| 230–239 | 88 | 0.08% |
| 240 | 10 | 0.01% |
| Threshold | Count Above | % of Sample |
|---|---|---|
| 65+ (clinical threshold) | 81,793 | 81.01% |
| 90+ | 68,943 | 68.28% |
| 120+ | 49,188 | 48.72% |
| 140+ | 34,153 | 33.83% |
| 160+ | 20,215 | 20.02% |
| 180+ | 9,194 | 9.11% |
| 200+ | 2,638 | 2.61% |
| 220+ | 400 | 0.40% |
| 240 (maximum) | 10 | 0.01% |
Standard RAADS-R reporting uses four broad clinical categories derived from the Ritvo validation study. These serve a clinical purpose but are too coarse to meaningfully reflect the range of experience within the above-threshold population — which in this dataset spans 175 score points from 65 to 240.
To address this, NeuroDirect has developed the NeuroDirect RAADS-R Bands: a seven-tier proprietary framework that maps score ranges onto descriptions grounded in how people in each range typically describe their experience.
n = 100,968 · NeuroDirect RAADS-R Bands © NeuroDirect
| Band | Score Range | Name | Count | % | Avg | SD |
|---|---|---|---|---|---|---|
| 1 | 0–64 | The Unlit Zone | 19,175 | 18.99% | 40.32 | 15.62 |
| 2 | 65–89 | The Threshold | 12,850 | 12.73% | 77.52 | 7.28 |
| 3 | 90–105 | The Recognisable | 10,131 | 10.03% | 97.58 | 4.64 |
| 4 | 106–119 | The Confirmed Pattern | 9,624 | 9.53% | 112.58 | 4.01 |
| 5 | 120–139 | The Clear Signal | 15,035 | 14.89% | 129.54 | 5.73 |
| 6 | 140–179 | The Deep Profile | 24,959 | 24.72% | 157.95 | 11.31 |
| 7 | 180–240 | The Full Spectrum | 9,194 | 9.11% | 194.48 | 11.89 |
Scores below the clinical threshold. Average 40.32, standard deviation 15.62. A below-threshold score does not rule out autism — masking, late-developing self-awareness, and atypical presentations can all produce scores below 65 in autistic individuals — but the result does not, on its own, support pursuing formal assessment on autism grounds.
Average 77.52, standard deviation 7.28 — the tightest clustering of any band. Autistic traits are present but may be subtle, highly context-dependent, or significantly masked. Many people in this band have developed sophisticated compensatory strategies across a lifetime. The cost of that masking — the exhaustion, the chronic sense of performing rather than being — is often invisible to those around them.
Average 97.58, std dev 4.64. A clear autistic profile is emerging across domains. People in this band often describe a pattern of lifelong difference — social confusion, sensory difficulties, intense interests — that has never had a name. The RAADS-R result at this level frequently prompts the response: "this is the first time I have seen my experience described."
Average 112.58, std dev 4.01 — the lowest standard deviation in the above-threshold bands, indicating a coherent and stable profile. Formal assessment is strongly indicated. People here frequently describe prior diagnoses of anxiety, depression, or ADHD recontextualised; lifelong social difficulty finally explained.
The most populated single band in the dataset (15,035 completions, 14.89%). Average 129.54, std dev 5.73. The mean (114.09) sits just below this band; the median (118) and mode (136) both fall within or adjacent to it. People scoring here typically describe autistic traits that have shaped every major domain of their life for as long as they can remember.
The largest band by count (24,959 completions, 24.72%). Average 157.95, std dev 11.31. Many people in this band describe a history of significant, cumulative difficulty in neurotypical environments — repeated job losses, relationship breakdowns, educational underachievement, misdiagnoses — without ever having had a framework to understand why.
9,194 completions, 9.11%. Average 194.48, std dev 11.89. The 400 people above 220, and the 10 who scored the maximum of 240, endorsed autistic traits at the ceiling of every dimension the scale measures. This is frequently the score of someone who has spent decades feeling not merely different but profoundly alien — often with multiple prior psychiatric diagnoses that never quite fit.
| Day | Completions | Avg Score | Std Dev |
|---|---|---|---|
| Monday | 13,884 | 114.71 | 49.21 |
| Tuesday | 14,240 | 114.81 | 49.23 |
| Wednesday | 14,727 | 115.13 | 48.64 |
| Thursday | 14,539 | 114.19 | 49.26 |
| Friday | 14,909 | 114.10 | 49.53 |
| Saturday | 14,917 | 111.89 | 50.04 |
| Sunday | 13,752 | 113.86 | 49.55 |
Scores are stable across all seven days — ranging only from 111.89 to 115.13 — consistent with the interpretation that people are responding based on stable, trait-level experience rather than transient mood.
The score distribution in this dataset is not what you would expect from a random population sample. A random general population sample would produce a distribution heavily weighted toward low scores — consistent with epidemiological estimates of autism prevalence at 1–2%. This distribution does the opposite: it rises through the 80s and 90s, peaks at 130–139, and tapers with a long right tail.
81.01% of people who completed this test scored above the clinical threshold. This is not a statement about general autism prevalence — it is a statement about the accuracy of self-directed help-seeking at scale.
No demographic data. We do not collect age, gender, ethnicity, or country of origin. We cannot report on how the score distribution varies by gender or age group.
No diagnostic status. We do not know how many of the 100,968 people hold a formal autism diagnosis or are in an assessment pathway.
Self-selected population. This distribution should not be used to estimate general population prevalence.
No repeat-completion control. We do not know how many individuals have completed the test more than once.
Self-administered context. The RAADS-R was developed for clinician-administered use. Self-administration introduces the possibility of response bias.
Ritvo et al. (2011) doi:10.1007/s10803-010-1133-5 · NeuroDirect dataset n=100,968 · Dashed line = clinical threshold (65)
Key score benchmarks compared:
Ritvo et al. (2011) · NeuroDirect n=100,968
Ritvo et al. (2011) — The original validation study. 779 participants across clinical sites in the US, Sweden, Australia, and the Netherlands. Mean score: 164.34 (autistic group), 64.59 (non-autistic controls). Sensitivity: 97%. Specificity: 100%.
Lundqvist & Lindner (2017) — Found that internet-administered RAADS-R produced results consistent with clinician-administered versions, supporting the validity of online self-administration.
Scale context — At 100,968 completions, the NeuroDirect dataset is approximately 130 times the size of the original Ritvo validation sample.
Community prevalence of self-identified autistic traits — The 81.01% above-threshold rate is a data point relevant to understanding who seeks out autism screening tools and why.
Online vs. clinician-administered RAADS-R — The consistency of this dataset's distribution provides naturalistic evidence relevant to questions about the reliability of self-administered versions.
The diagnostic access gap — The scale of unmet diagnostic need becomes quantifiable rather than theoretical.
Distribution norms for community samples — The percentile table in Section 3.4 may be useful to researchers working with community-administered RAADS-R data.
Longitudinal stability — The consistency of mean scores between earlier (n=72,111, mean=111.51) and current (n=100,968, mean=114.09) snapshots suggests distributional stability over time.
We are actively seeking research collaborators. Please contact us to discuss data access or a collaborative study.
The 34,153 people who scored above 140 — one third of all completions — are in a score range that clinical guidance consistently associates with a pronounced autistic profile. Most of them are not currently in an assessment pathway.
NeuroDirect hosts a directory of neurodivergence specialists across the UK. List your practice here to reach a population that has already demonstrated motivation at every step. Or contact us to discuss partnership opportunities.
Please contact us to discuss.
Takes around 20 minutes. Results available immediately. Not a diagnosis — but for many people, the beginning of one.
In this dataset of 100,968 community completions, the mean RAADS-R score was 114.09, the median was 118, and the most common individual score was 136.
65 is the validated clinical threshold above which results are considered consistent with an autistic profile. In this dataset, 81.01% of completions scored at or above 65.
In this community sample, the 90th percentile is 175 and the 99th percentile is 203. A score of 140 or above places a respondent in the top third of this dataset — The Deep Profile band.
The validated threshold is 65. Many clinicians consider scores above 90, 120, or 140 as progressively stronger indicators. In this dataset, 33.83% scored above 140 and 9.11% scored above 180.
Lundqvist & Lindner (2017) found that online self-administration produced results consistent with clinician-administered versions. The stability of scores across the growth of this sample is consistent with reliable self-administration.
No. An elevated score is a strong indicator that formal assessment is warranted — not a diagnosis in itself. If your score is above 65, seek assessment from a qualified professional.
The NeuroDirect RAADS-R Bands are a proprietary seven-tier framework providing more granular context for RAADS-R results than the standard four clinical categories. Full band statistics are in Section 5.2.
See the citation section below.
Ritvo, R. A., Ritvo, E. R., Guthrie, D., Ritvo, M. J., Hufnagel, D. H., McMahon, W., Tonge, B., Mataix-Cols, D., Jassi, A., Attwood, T., & Eloff, J. (2011). The Ritvo Autism Asperger Diagnostic Scale-Revised (RAADS-R): A scale to assist the diagnosis of autism spectrum disorder in adults. Journal of Autism and Developmental Disorders, 41(8), 1076–1089. https://doi.org/10.1007/s10803-010-1133-5
Lundqvist, L.-O., & Lindner, H. (2017). Is the Ritvo Autism and Asperger Diagnostic Scale-Revised (RAADS-R) a useful diagnostic instrument for adults with autism spectrum disorders in a clinical setting? Advances in Neurodevelopmental Disorders, 1(3), 128–135. https://doi.org/10.1007/s41252-017-0021-y
Data sourced from 100,968 completed RAADS-R test sessions on NeuroDirect.co.uk. All zero scores excluded. Internal operator completions excluded. NeuroDirect does not collect demographic data at test completion. The NeuroDirect RAADS-R Bands are a proprietary framework and do not replace clinical interpretation. This article is for informational purposes only and does not constitute clinical advice. Individuals with elevated scores are strongly encouraged to seek formal assessment from a qualified professional. Dataset version: July 2026.