What the CARS-2 Measures and What It Misses

What the CARS-2 Actually Rates

The scale's structure comes from the original Childhood Autism Rating Scale, published in 1980: fourteen areas of behavior plus a fifteenth rating of general impressions of autism, each scored from 1, meaning within normal limits, to 4, meaning severely abnormal, in half-point steps, giving a total somewhere between 15 and 60. The second edition's standard version carries that structure forward 1Ref 1Schopler E, Reichler RJ, DeVellis RF, Daly K (1980).Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS).The scale's 15-item structure (fourteen behavioral areas plus a fifteenth general-impressions rating), per-item scoring from 1 to 4 in half-point increments, the 15-to-60 total range, the direction of scoring in which higher means greater impairment, and that this original CARS is the structural basis of the CARS-2 standard version. Not cited for cutoff bands, which belong to the publisher's manual..

Higher totals mean greater impairment on this scale, not greater ability 1Ref 1Schopler E, Reichler RJ, DeVellis RF, Daly K (1980).Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS).The scale's 15-item structure (fourteen behavioral areas plus a fifteenth general-impressions rating), per-item scoring from 1 to 4 in half-point increments, the 15-to-60 total range, the direction of scoring in which higher means greater impairment, and that this original CARS is the structural basis of the CARS-2 standard version. Not cited for cutoff bands, which belong to the publisher's manual.. That direction is worth fixing in mind before reading a report, because plenty of other tests in the same file run the opposite way — cognitive and adaptive measures usually put higher scores on the side of more skill. Two numbers on facing pages can be inverted without anyone noticing.

The half-point steps matter more than they look. They exist because real children sit between the written descriptions, and forcing a rater to pick one whole point or the next would throw away information they genuinely have. It is a sign the scale was built by people watching children rather than by people designing a form.

The second edition comes in more than one form, including a version intended for people with stronger verbal and cognitive skills and a questionnaire completed by parents or caregivers. Which form produced the number on a report is a fair and useful question to ask, because they are not interchangeable and the report does not always say.

What the Total Means, and What It Does Not

A CARS-2 total describes how far a rater judged a child's behavior to sit from typical development across fifteen defined areas, on a scale where higher means more impairment 1Ref 1Schopler E, Reichler RJ, DeVellis RF, Daly K (1980).Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS).The scale's 15-item structure (fourteen behavioral areas plus a fifteenth general-impressions rating), per-item scoring from 1 to 4 in half-point increments, the 15-to-60 total range, the direction of scoring in which higher means greater impairment, and that this original CARS is the structural basis of the CARS-2 standard version. Not cited for cutoff bands, which belong to the publisher's manual.. That is a description of ratings. It is not a measurement of a child the way a hemoglobin level measures blood, and no single medical test diagnoses autism in the first place 2Ref 2Centers for Disease Control and Prevention (2024).About Autism Spectrum Disorder.That autism is a developmental disability in which people may communicate, interact, behave, and learn differently, and that there is no single medical test that diagnoses it..

The word that causes the most damage is severe. On this scale it summarizes how one rater coded fifteen items in a defined window. It is not a statement about intelligence, not a prediction of what a child will manage at ten or twenty, and not a ranking of how much a child is suffering. Autism is a developmental disability in which people communicate, interact, behave, and learn differently 2Ref 2Centers for Disease Control and Prevention (2024).About Autism Spectrum Disorder.That autism is a developmental disability in which people may communicate, interact, behave, and learn differently, and that there is no single medical test that diagnoses it., and differently is the load-bearing word there. A rating scale is a rough instrument for it.

Rating scales as adjuncts is the accurate frame. They are structured aids that make one clinician's observations comparable and reviewable, sitting beside developmental history, direct observation, language and cognitive testing, and what school reports. Any one of those can be wrong alone. That is not a defect in the scale; it is why evaluations use several.

Why This Page Does Not Print the Score Bands

The CARS cutoff bands are findable online in about ten seconds. This page will not print them, for two reasons, and squeamishness is not either of them. The bands usually quoted trace to the test's own manual rather than to the 1980 paper that established the scale's items and scoring 1Ref 1Schopler E, Reichler RJ, DeVellis RF, Daly K (1980).Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS).The scale's 15-item structure (fourteen behavioral areas plus a fifteenth general-impressions rating), per-item scoring from 1 to 4 in half-point increments, the 15-to-60 total range, the direction of scoring in which higher means greater impairment, and that this original CARS is the structural basis of the CARS-2 standard version. Not cited for cutoff bands, which belong to the publisher's manual., and the second edition's norms and version-specific thresholds sit in a publisher's manual that is not a peer-reviewed source at all.

The deeper reason is what a band does once it is loose. A number becomes a band, a band becomes a label, and the label lands on a child whose ratings you did not make, from written descriptions you have never seen, on a scale you were not trained to use. A total on someone else's rating form is not something a parent is supposed to be able to interpret alone. Not being able to is the normal case, not a gap in you.

What is worth asking, and what a competent evaluator will answer without hesitation: which form of the scale was used, what the ratings were based on, which areas pushed the total up, and what that suggests about where support would help most. Those answers are usable. The number by itself mostly is not.

What One Session Misses

The CARS-2 rates what a rater can see and learn about, so it inherits every limit of that. A child in an unfamiliar room with an unfamiliar adult may show less than usual, or considerably more. A child having a good hour rates differently from the same child having a bad one. And a child working hard to hold themselves together gets rated on the holding-together, not on what it cost.

Masking is the name families eventually learn for that last one. People who have learned to copy social behavior — more often, though not only, girls and older speaking children — can present in a clinic in ways a behavior-rating scale reads as unremarkable, while everything is expensive at home afterward. A scale built on observable behavior has no column for effort.

Identification patterns show the same limitation at population scale. CDC surveillance finds that measured autism prevalence varies widely between communities and reads that variation as reflecting differences in identification and access to services rather than real differences in how often autism occurs 3Ref 3Centers for Disease Control and Prevention (2025).Autism Prevalence Varies Across US Communities.That measured autism prevalence varies widely between communities and that this variation reflects differences in identification and access to services rather than necessarily true differences in occurrence.. Who gets evaluated, by whom, and with how much time is part of what any rating reflects.

Co-occurring conditions blur the picture further. Anxiety, attention problems, a language disorder, hearing loss, and chronic poor sleep all change how a child presents during a rating session, and a single total absorbs all of it without saying which part came from where.

The CARS-2 Next to a Questionnaire Like the GARS-3

Two instruments a family may see both described as an autism rating scale work quite differently. The CARS-2 is rated by an observer from what they gathered, producing a total on a fixed 15-to-60 range 1Ref 1Schopler E, Reichler RJ, DeVellis RF, Daly K (1980).Toward objective classification of childhood autism: Childhood Autism Rating Scale (CARS).The scale's 15-item structure (fourteen behavioral areas plus a fifteenth general-impressions rating), per-item scoring from 1 to 4 in half-point increments, the 15-to-60 total range, the direction of scoring in which higher means greater impairment, and that this original CARS is the structural basis of the CARS-2 standard version. Not cited for cutoff bands, which belong to the publisher's manual.. The GARS-3 rating scale is a norm-referenced questionnaire: 58 items across six subscales, for people aged 3 to 22, reported as subscale scaled scores and an overall Autism Index, with higher scores again indicating greater likelihood of autism 4Ref 4Karren BC (2017).A Test Review: Gilliam, J. E. (2014). Gilliam Autism Rating Scale–Third Edition (GARS-3).That the GARS-3 is a norm-referenced instrument of 58 items across six subscales for ages 3 to 22, reported as subscale scaled scores and an Autism Index with higher scores indicating greater likelihood of ASD; and the review's stated cautions about ages 20 to 22, minority populations, and a norming sample composed only of individuals with ASD diagnoses with limited racial and ethnic diversity. Not cited for cutoffs or accuracy figures..

Norm-referenced is the real difference. A norm-referenced score says where someone sits relative to the sample the test was built on, which makes the composition of that sample part of the answer. The peer-reviewed review of record says so directly, recommending caution when assessing people aged 20 to 22 and people from minority populations, and noting that the norming sample consisted only of individuals already diagnosed with autism, with limited racial and ethnic diversity 4Ref 4Karren BC (2017).A Test Review: Gilliam, J. E. (2014). Gilliam Autism Rating Scale–Third Edition (GARS-3).That the GARS-3 is a norm-referenced instrument of 58 items across six subscales for ages 3 to 22, reported as subscale scaled scores and an Autism Index with higher scores indicating greater likelihood of ASD; and the review's stated cautions about ages 20 to 22, minority populations, and a norming sample composed only of individuals with ASD diagnoses with limited racial and ethnic diversity. Not cited for cutoffs or accuracy figures..

Neither type replaces the other, and neither is a diagnosis. If both appear in one report and appear to disagree, that is not automatically an error. They ask related questions in different ways, of different raters, against different reference points.

How a Child Gets to the Point of Being Rated

Nothing about a rating scale happens first. At the front of the path is a pediatric clinician doing ongoing developmental surveillance at routine visits, adding formal screening at recommended ages, and referring onward when a concern surfaces 5Ref 5Centers for Disease Control and Prevention (2024).Information on Autism Spectrum Disorder for Healthcare Providers.That the primary-care role in early identification combines ongoing developmental surveillance with formal screening and referral when concerns arise, placing a diagnostic rating scale downstream of that pathway.. A CARS-2 sits well downstream of that, inside a diagnostic evaluation a family has already been referred into.

The stretch between referral and evaluation is where most families actually spend their year, and it is rarely short. Two routes sometimes shorten it: a university training clinic, where supervised graduate clinicians evaluate at reduced cost and often against a different waitlist, and a telehealth evaluation, where a remote practice can sometimes see a family months earlier than a local one.

Neither is second-best by definition, and neither suits every child. What matters is that whoever does the evaluation is qualified to diagnose in your state and produces a report your school district and your health plan will actually accept.

What Happens to the Number Afterward

A CARS-2 total does not stay in the evaluation room. It goes into a report, and that report gets read by people deciding about services: a school team, a health plan reviewing a therapy request, sometimes a state early-intervention program. Autism therapy coverage is common across payers but governed by plan-specific rules on documentation, medical necessity, and coding 6Ref 6American Speech-Language-Hearing Association (2024).Payer Portal: Autism Spectrum Disorder.That autism-related therapies are commonly covered but subject to payer-specific coverage and coding rules, used here for the general statement without any specific benefit amounts..

That has one practical consequence worth knowing in advance. When a report is going to be read by someone deciding whether to authorize services, the parts that carry weight are usually the concrete descriptions of what a child can and cannot do, not the headline total. Families who ask their evaluator to be specific — what mealtimes look like, what happens at transitions, what a full school day costs this particular child — tend to end up with reports that hold up better in front of a stranger.

And a total is a snapshot with a date on it. It describes ratings made on one day about one window of behavior. Children change; a filed report does not.