There is no single IQ for nonverbal autistic people, and the number a child comes home with depends enormously on which test was used. Autistic people have been found to score around 30 percentile points higher on a nonverbal reasoning test than on a standard IQ battery, a gap not seen in non-autistic people. If your child is non-speaking and a report came back with a low number, the test is part of that story.
If your child is young and you are at the start of this, early intervention ABA is usually where the conversation begins, and it does not require you to have the assessment picture sorted out first.
Why There Is No Single Number
An autism diagnosis says nothing about intelligence on its own. The DSM-5 criteria published by the CDC describe differences in social communication and restricted or repetitive behavior. Intellectual ability is recorded separately, as a specifier, precisely because it varies so widely across autistic people.
Non-speaking is also not one thing. Some children have no spoken words. Some have a handful. Some speak in some settings and not others. Researchers increasingly use "minimally verbal" or "non-speaking" rather than "nonverbal," because the older term implies an absence of communication rather than an absence of speech, and those are different.
What the Research Actually Shows
A 2025 study in Frontiers in Psychiatry compared minimally verbal autistic children with verbal autistic peers. On nonverbal measures, the minimally verbal group averaged 59.6 and the verbal group averaged 87.6.
That is the group average, and it is lower. Two things sit alongside it that matter more for an individual family.
The first is that 26.5% of the minimally verbal children scored at or above 70 on nonverbal measures, in the typical or borderline range. More than one child in four. The authors described this as a fundamental divergence between verbal and nonverbal functioning, and noted it contradicts the assumption that reduced language necessarily means compromised intellect.
The second is that a group average cannot tell you anything about your child. A mean of 59.6 is compatible with a child scoring 40 and a child scoring 95 sitting in the same study.
The Test Changes the Score
This is the finding most worth knowing. Work by Dawson, Soulières, Gernsbacher, and Mottron found that autistic people scored roughly 30 percentile points higher on Raven's Progressive Matrices than on Wechsler scales, while non-autistic people showed no such gap. A follow-up study in PLOS ONE found the same pattern across the spectrum, with a 27 percentile point advantage in one adult group.
Thirty percentile points is not a rounding error. It is the difference between a report that says "well below average" and one that says "within normal limits," for the same person on the same day.
Why Standard IQ Tests Misread Non-Speaking Children
A conventional IQ battery asks a child to do several things at once that have nothing to do with reasoning:
- Understand spoken instructions from an unfamiliar adult
- Produce a spoken or precisely pointed answer
- Do it within a time limit
- Sustain attention in an unfamiliar room
- Care enough about a stranger's approval to keep trying
A child who reasons perfectly well but cannot do the second item will score low. The test has measured the speech barrier, not the thinking behind it.
Which Assessments Are Better Suited
Nonverbal measures such as the Leiter scales or Raven's Matrices reduce the spoken-language demand and are commonly used when speech is limited. They are not perfect either, since they still require attention, compliance, and a motor response.
What usually helps most is not a different single test but a different approach: several sources of information, an assessor who has worked with non-speaking children, and a response method the child already uses reliably, whether that is pointing, eye gaze, picture exchange, or a speech-generating device.
A Worked Example from a Session
Here is the shape this often takes.
A child arrives with a report placing them well below average, based on an assessment that required spoken answers. Nobody is re-testing anything in session. The team is watching what happens when the response requirement changes.
Given an array of pictures and a way to indicate a choice, by touching, by looking, or by handing over a card, you start to see what the child can sort, match, sequence, and anticipate. The team notes which categories the child already has, which instructions land when delivered visually rather than verbally, and how long attention holds when the task is not a spoken quiz.
None of that is an IQ score, and it is not meant to be. It is a record of what the child demonstrably knows, documented well enough that the next evaluator has something to work from. Our post on skill acquisition programs covers how those observations become teaching targets.
What Parents Should Do Next
- Ask which instrument was used and how much spoken language it required. That one question explains a lot of low scores.
- Ask what response method was accepted. If your child communicates by pointing or device and the test required speech, the result reflects the mismatch.
- Get a reliable response channel in place first. Any assessment is more accurate once a child has a dependable way to answer. Written and drawn output is another route, covered in our post on ABA and writing skills.
- Keep your own evidence. Short videos of your child solving something, following a routine, or making a choice are concrete and hard to argue with.
- Do not plan around the number. Teaching targets should come from what your child can and cannot currently do, not from a score. Our post on ABA therapy effectiveness covers what the evidence supports targeting.
Your Child's Right to a Fair Assessment in North Carolina and Virginia
If a school evaluation produced the low number, both states give you specific ground to stand on.
Virginia's regulations require that a variety of assessment tools be used and that "no single measure or assessment is used as the sole criterion." Assessments must be administered "in the form most likely to yield accurate information on what the child knows and can do," and tools must be selected so that results reflect aptitude "rather than reflecting the child's impaired sensory, motor, or communication skills."
That last clause is the whole argument. A verbal-loaded test given to a non-speaking child measures the communication impairment, which is exactly what the regulation says it must not do.
North Carolina's Policies Governing Services for Children with Disabilities are similar. Eligibility "must be determined by using multiple sources of data and must not be dependent upon single test scores," drawing on observations, interviews, play assessment, adaptive and developmental scales, and clinical judgment. The policies also state that where standardized instruments are "clearly inappropriate as standardized," the psychologist should apply professional judgment.
So the question to put to a school team, in writing, is which measures were used, what response method each required, and what else was gathered beyond the test score. Families we work with in Charlotte often find that asking it in writing changes what comes back.
Conclusion
There is no fixed IQ for nonverbal autistic people. Group averages for minimally verbal children sit lower than for verbal autistic children, and more than a quarter still score in the typical or borderline range on nonverbal measures. The instrument used can move a score by around 30 percentile points.
What follows is practical rather than reassuring. Ask what was used and what it demanded of your child, get a dependable response channel in place, keep your own evidence, and build teaching around what your child demonstrably does rather than a number on a page.
Ready to Take the Next Step?
Connect n Care ABA provides in-home ABA therapy, school-based ABA, and ABA parent training to families across North Carolina and Virginia.
If a report does not match the child you know, contact our team, and we will talk through what to ask for next.
SOURCES:
- https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0025372
- https://www.frontiersin.org/journals/psychiatry/articles/10.3389/fpsyt.2025.1549092/full
- https://www.cdc.gov/autism/hcp/diagnosis/index.html
- https://www.cdc.gov/autism/signs-symptoms/index.html
- https://law.lis.virginia.gov/admincode/title8/agency20/chapter81/section70/
- https://www.dpi.nc.gov/documents/publications/catalog/ec144-policies-governing-services/open
