top of page
Search

Childhood Autism Rating Scale Explained: CARS vs. CARS-2, Cutoffs, and Use Cases

  • Writer: Jamie P
    Jamie P
  • Aug 28
  • 7 min read
ree

A clear, clinic-ready guide to the Childhood Autism Rating Scale—how CARS and CARS-2 differ, what the cutoffs mean, who should use which form (ST vs. HF), and how to report results responsibly across schools, clinics, and research.


The Childhood Autism Rating Scale (CARS) is one of the most widely used clinician rating tools for identifying autism characteristics and describing severity. It’s valued because it’s fast, systematic, and observation-anchored. But it’s also easy to misunderstand—especially now that CARS-2 offers multiple forms, nuanced cutoffs, and updated guidance.


This guide demystifies the scale. You’ll learn exactly what CARS measures, how scoring and severity bands work, how CARS-2 Standard (ST) differs from CARS-2 High-Functioning (HF) and the Questionnaire for Parents or Caregivers (QPC), and how to use results ethically—never as a stand-alone diagnosis, always as part of a comprehensive evaluation.


What CARS Measures and Why It’s Useful

CARS is a clinician rating scale with 15 items covering areas that tend to differ for autistic individuals—such as relating to people, emotional response, body use, adaptation to change, communication, and sensory responses. Each item is scored along a continuum based on direct observation and corroborated history, yielding a total score that reflects the degree of autistic features observed during the assessment period.


Why teams choose CARS/CARS-2:

  • Efficiency: It can be completed quickly after sound observation and record review.

  • Breadth: Items span social-communication, behavior, and sensory domains.

  • Flexibility: Appropriate across settings (clinics, schools, research).

  • Communication: Produces a single severity estimate that’s easy for non-specialists to grasp.


Good practice reminder: CARS does not replace a comprehensive diagnostic evaluation. It’s a structured judgment—one piece of a larger puzzle that includes developmental history, caregiver interview, speech-language and cognitive testing as indicated, and standardized autism observations.


How CARS Scoring Works

CARS items are rated on a 1 to 4 scale (with half-point options) to reflect how characteristic or intense the observed behavior is relative to developmental expectations. Summing the 15 items yields a total score from 15 to 60:

  • 15–29.5: Within non-autistic range

  • 30–36.5: Mild to Moderate autism characteristics

  • 37–60: Severe autism characteristics

These classic bands come from the original CARS research and remain widely cited. Keep in mind that severity on CARS is not identical to clinical “severity levels” used for service planning; it is a rating-scale severity intended to summarize what was observed.


CARS vs. CARS-2: What Changed

CARS-2 keeps the core 15-item structure but introduces forms tailored to age, language, and cognitive profile, improving fit and clinical accuracy:

  • CARS-2 Standard (CARS2-ST): Best for children under 6 or individuals of any age with notable communication delays or below-average cognitive abilities.

  • CARS-2 High-Functioning (CARS2-HF): Designed for age 6+ who are verbally fluent and typically have average or higher cognitive skills (often described as IQ ≈ 80+ in studies). The HF form emphasizes items that better differentiate autistic traits when language is stronger and adaptive skills are higher.

  • CARS-2 Questionnaire for Parents or Caregivers (CARS2-QPC): A structured informant form to support clinician ratings; it is not a stand-alone diagnostic instrument.


Why this matters: using the wrong form can inflate or deflate scores. A verbally fluent teen might look “less autistic” on the ST form because some items aren’t sensitive at the higher end, whereas the HF form targets differences that still matter when language and general reasoning are stronger.


Cutoffs and Interpretation Including Newer Findings

Historically, a total score ≥30 has been used as a cutoff suggesting autism-consistent features (with 30–36.5 in the mild-to-moderate band and ≥37 in the severe band). With CARS-2, research has explored optimized thresholds—particularly for the ST form. Some studies suggest that scores in the high 20s (e.g., ~28.5) may flag children who meet broader “autism spectrum” criteria compared with gold-standard observational tools.


What to do in practice:

  • Use publisher guidance and your local norms first.

  • For the ST form, know that some teams apply ≥30 as a conservative “autism likely” cutoff and consider 28.5–29.5 a “watch” range when converging evidence (history/ADOS-style observation) points toward autism.

  • For the HF form, follow the manual’s banding and your program’s policy; HF cutoffs can differ from ST because the item content differs.


Key principle: Never over-index on a single threshold. Confirm with a comprehensive evaluation, especially for borderline scores.


Use Cases: When CARS-2 Shines and When It Doesn’t


Best-Fit Situations

  • Clinician-observed evaluations where you need a quantified severity estimate grounded in real-time behavior.

  • School-based multidisciplinary assessments that combine observation, teacher and caregiver input, and standardized measures to inform eligibility and supports.

  • Research and program evaluation seeking a continuous severity metric to track differences across cohorts or over time.


Situations That Need Caution

  • Brief telehealth only: CARS was developed for direct observation. While telehealth can supplement history gathering, rating exclusively from video calls risks missing subtle social-communication cues and behavior patterns.

  • Screening in primary care: For front-door screening, use pediatric screeners like M-CHAT-R/F; CARS/CARS-2 is better as part of the diagnostic evaluation step.

  • Progress monitoring without context: A lower score later may reflect adaptation to the testing context, not necessarily change in underlying differences. Pair CARS with goal-level measures and functional outcomes.


CARS-2 ST vs. CARS-2 HF: Choosing the Right Form

Start with language and cognitive profile, then age:

  • Under 6 or significant language/cognitive delays at any age → ST

  • Age 6+, verbally fluent, average-range cognition → HF


Practical tips:

  • If you’re unsure, pilot both (using QPC input and your observation notes) and see which item set better captures meaningful differences. You’ll still report one official form in the record.

  • Keep a local decision tree so your team applies forms consistently across clinicians and sites.

  • For bilingual or multilingual families, plan interpreting support for the caregiver questionnaire and interview to prevent under- or over-identification due to language mismatch.



CARS-2 and Other Tools (ADOS-2, M-CHAT-R/F): How They Fit Together

  • M-CHAT-R/F is a screening tool for toddlers in primary care; it flags risk and prompts referral.

  • ADOS-2 is a standardized observational assessment with structured activities and algorithm cutoffs.

  • CARS-2 is a clinician rating scale that synthesizes observation and history into a continuous severity score.


A common workflow:

  1. Screen in primary care (M-CHAT-R/F) →

  2. Comprehensive evaluation with clinical interview, developmental history, language/cognitive testing as appropriate →

  3. Direct observation (often ADOS-style) + CARS-2 rating to summarize severity and support the diagnostic formulation.

Using several converging sources strengthens clinical confidence—especially near cutoffs.


Administration: Training, Ethics, and Time


Who administers? 

Trained clinicians (e.g., psychologists, developmental-behavioral pediatricians, experienced school teams).


What’s required?

  • Direct observation in a naturalistic or structured context.

  • Caregiver interview/QPC as supplemental evidence—never as the only source.

  • Enough time to see social communication, play/problem-solving, transitions, and sensory responses (often during a broader evaluation session).


Ethics and copyright

CARS-2 is a proprietary instrument. Do not reproduce test items, scoring rubrics, or forms in reports. Summarize clinically and cite the instrument properly.



Writing Clear Reports With CARS-2

Strong reports avoid data-dumping and speak to practical supports. Consider this structure:

  1. Reason for referral & methods (history sources, observation contexts, instruments used).

  2. CARS-2 form used (ST or HF) and the rationale for choosing it.

  3. Behavioral evidence item-by-item in prose (no copyrighted item text), with examples that illustrate ratings.

  4. Total score and interpretive band (e.g., falls in “mild to moderate” range).

  5. Integration with other findings (language, cognition, adaptive behavior, mental health).

  6. Clinical formulation (diagnostic statement if warranted, differential diagnosis considered, and functional profile).

  7. Recommendations (communication supports, school accommodations, referral to community resources, caregiver training, and follow-up plan).

Keep language family-friendly. Explain that a CARS-2 score summarizes observed differences, not a person’s potential, and that supports are individualized.


Common Pitfalls and Easy Fixes

  • Using the wrong form → Build a one-page form-selection decision tree and train to it.

  • Treating CARS-2 as diagnostic proof → Make it explicit that diagnosis rests on converging data.

  • Over-relying on caregiver questionnaire → Use QPC as input; base ratings on observation plus interview.

  • Ignoring language/culture → Offer interpreting, provide translated materials, and consider cultural norms when interpreting social behaviors.

  • Comparing ST and HF scores directly → They are not interchangeable; document the form and context before comparing across time.


Equity, Language Access, and Context

Autism looks different across cultures and languages, and some behaviors are valued or expected differently by community norms (eye contact, play styles, conversational pacing). That doesn’t make CARS-2 unusable; it makes context essential. Do this:

  • Involve interpreters trained in clinical settings.

  • Gather examples from caregivers and educators to anchor ratings in the person’s everyday environment.

  • Consider co-occurring conditions (ADHD, anxiety, language disorder, intellectual disability) that influence observed behavior in session.


Telehealth and Remote Observation: What’s Reasonable

Telehealth is powerful for interviews and follow-up, but CARS-2 is built around direct observation. If you must use video:

  • Ask families to position the camera to capture face orientation, gestures, and joint attention.

  • Avoid laggy platforms; disruptions bias ratings.

  • Treat telehealth ratings as provisional if key behaviors weren’t observable; complete in-person observation when feasible.



Implementation Checklist for Clinics and Schools

Before the Visit:

  • Confirm form selection (ST vs. HF) from intake data.

  • Send QPC in the family’s preferred language with clear return instructions.

  • Prepare an observation plan (play, conversation, tasks, transitions) that surfaces social-communication and behavior patterns.


During the Visit:

  • Observe unstructured and structured interactions.

  • Note specific examples for each domain (e.g., response to name, pretend play, flexibility with change).

  • Capture sensory responses and regulation strategies.


After the Visit:

  • Integrate history + observation into ratings.

  • Double-check cutoff interpretation for the selected form.

  • Write a family-friendly summary and offer next steps (services, school IEP/504 process, community supports).


CARS-2 in Research and Quality Improvement

Because CARS-2 yields a continuous severity score, it’s widely used to compare cohorts and track outcomes in programs. If you use it for research or QI:

  • Keep form selection constant across timepoints or stratify analyses (ST vs. HF).

  • Report observation context (clinic vs. school; in-person vs. telehealth).

  • Pair CARS-2 with functional outcomes families care about (communication, participation, adaptive skills) rather than severity alone.


Frequently Asked Questions

  • Is CARS-2 a replacement for ADOS-2?

    No. ADOS-2 is a structured observational assessment; CARS-2 is a clinician rating scale. Many teams use both to triangulate findings.

  • Can schools use CARS-2 for eligibility decisions?

    Yes—as part of a multidisciplinary evaluation that includes observation, interviews, and other standardized measures. Don’t base eligibility solely on a CARS-2 score.

  • Do CARS-2 scores equal clinical “Level 1/2/3” severity?

    No. CARS-2 provides a rating-scale severity estimate. Clinical severity levels reflect support needs across settings and should be informed by adaptive behavior, language, and functional impact.

  • What if a child scores 29–30?

    Borderline scores require more data, not less. Review history, repeat observation, or add a structured observation (e.g., ADOS-style tasks) before concluding.


About OpsArmy

OpsArmy builds AI-native back-office operations as a service (OaaS). We help clinics and schools run smoother with trained, managed teams who support scheduling, intake, documentation, and coordination—so clinicians and educators can focus on people, not paperwork. 



Sources


 
 
 

Comments

bottom of page