Our method

Updated September 12, 2026. No clinician has reviewed this page.

This page exists so nothing about how our screenings work has to be taken on faith. It lists the four instruments we offer, exactly how each one is scored in our code, where our thresholds match the original published research, where they differ, what none of the four can measure, and how your data is handled while you take them.

Jack Squire built this site and is not a clinician. No clinician has reviewed the scoring, the wording, or any page on this site. Every number below is either drawn directly from the cited research or clearly marked as our own addition.

AQ-10

The AQ-10 was published in 2012 by Carrie Allison, Bonnie Auyeung, and Simon Baron-Cohen, drawn from the 50-item Autism-Spectrum Quotient. Our version presents all 10 items on a four-point scale from Definitely Agree to Definitely Disagree. Four items are scored in the agree direction (choosing Definitely or Slightly Agree earns a point) and six are scored in the disagree direction (choosing Slightly or Definitely Disagree earns a point). Your total runs from 0 to 10, and we mark 6 or above as meeting our threshold. This is the same threshold the original 2012 paper recommends for referral, and the same one NICE guidance uses in the UK, so our implementation and the published research agree on the AQ-10 without any adjustment on our part.

AQ-50

The AQ-50 was published in 2001 by Simon Baron-Cohen, Sally Wheelwright, Richard Skinner, Joanne Martin, and Emma Clubley. Our version presents all 50 items on the same four-point agreement scale used on the AQ-10, split into five domains of 10 questions each: Social Skills, Attention Switching, Attention to Detail, Communication, and Imagination. Each qualifying answer earns one point, for a total between 0 and 50, with each domain separately scored out of 10.

Here our threshold differs from the original research, and we want to say that plainly rather than let it hide in the fine print. The original 2001 paper used a score of 32 or above as its single threshold for clinically significant traits. Our tool flags a score of 26 or above as meeting our own threshold, a lower bar than the published study uses, and separately treats 32 or above as an additional higher band that lines up with the original research. We set 26 as an earlier flag so that people whose scores fall in the middle range are told their results are worth a closer look, rather than being sorted into a strict pass or fail against the original study's number alone. If you want to compare your AQ-50 result strictly against the published research, use 32; if you want to know when our tool starts calling a result worth attention, that number is 26.

RAADS-R

The RAADS-R was published in 2011 by Riva Ariella Ritvo and colleagues, designed specifically for adults. Our version uses the same four-point lifetime pattern scale as the original: "True now and when I was young" scores 3, "True only now, not when I was young" scores 2, "True only when I was younger than 16" scores 1, and "Never true" scores 0. A number of items are reverse-scored before being added to your total, because the statement itself describes a non-autistic pattern. Your total runs from 0 to 240 across all 80 items, split across four domains: Language, Social Relatedness, Sensory-Motor, and Circumscribed Interests. We mark a total of 65 or above as meeting the threshold, which is the same cutoff used in the original 2011 paper, so our implementation and the published research agree here as well. We additionally describe scores of 130 or above with stronger language in your results, but that second line is our own way of flagging a score well above threshold; the original paper does not draw that specific second distinction.

M-CHAT-R/F

The Modified Checklist for Autism in Toddlers, Revised, with Follow-Up, was validated by Diana Robins and colleagues and published in 2014. It is a parent-reported screening for children between 16 and 30 months, using 20 yes or no questions about the child's behavior. Our version scores each answer that indicates developmental risk as one point, for a total between 0 and 20, and reports three bands: Low Risk (0 to 2), Medium Risk (3 to 7), and High Risk (8 to 20). This matches the published low, medium, and high risk structure of the instrument. One limitation worth stating clearly: the full published M-CHAT-R/F protocol includes a structured Follow-Up interview for children who score in the medium range, with additional scripted questions designed to clarify borderline answers before a referral decision is made. Our tool implements the 20-question first-stage checklist only; it does not implement that Follow-Up interview, so a Medium Risk result here should be treated as a prompt to pursue that follow-up step with a pediatrician, not as a finished assessment on its own.

What none of these screenings measure

• Developmental history. A formal autism diagnosis looks at how traits have shown up since childhood and how much they affect functioning across settings. A questionnaire answered in one sitting cannot gather that history.
• Functional impairment. None of these tools measure how much support a person actually needs day to day. Two people can land on the same score while living very different lives.
• Differential diagnosis. Anxiety, ADHD, OCD, PTSD, and several other conditions can raise scores on some items on every one of these instruments without the person being autistic. Only a clinician weighing the full picture can sort that out.
• Masking. All four tools rely on self-report (or, for the M-CHAT-R/F, parent report), which means they can only measure what the respondent is able to observe and willing to describe. Years of masking can suppress the very patterns these screenings are trying to detect.

Browser-side scoring

Every calculation described above runs in your browser. When you answer a question, that answer is held in memory on your device and used to compute your score locally; your individual answers are not transmitted to our servers as part of taking a screening. If you choose to enter your email to receive your results or join our list, that is a separate, explicit action, and only the summary result you choose to share (such as your band, not your item-by-item answers) is used for that purpose. No account is required to take any screening on this site.

No clinician has reviewed this site

We want this stated somewhere unambiguous rather than folded into a footer disclaimer. Jack Squire, founder of this site, is not a clinician. No clinician has reviewed the scoring logic, the interpretive text, or any page describing these instruments. Every fact and figure on this page is drawn directly from the cited, peer-reviewed research or is explicitly marked as our own addition, as with the AQ-50 and RAADS-R bands described above. If you find an error, see our editorial policy for how to report it.

Contact

Questions about how a specific screening is scored, or corrections to anything on this page, can be sent through the contact details listed on our About page. We read every message and correct factual errors when they are identified.