Validity and reliability of the Childhood Autism Rating Scale with autistic adolescents.
CARS is still a solid autism screener for teens—just delete the one item that drags the total score down.
01Research in Context
What this study did
The team tested if the Childhood Autism Rating Scale works for teens. They gave CARS to 45 autistic and 41 non-autistic adolescents. All kids were 12 to 18 years old.
They checked if scores could tell the groups apart. They also looked at each item to see if any hurt the total score.
What they found
CARS scores were much higher for the autistic group. One item, "Level and consistency of intellectual response," dragged the total down. Dropping it made the scale stronger.
The scale stayed reliable and valid for teens, just like it does for younger kids.
How this fits with other research
Matson et al. (1994) later built on this work. They found three CARS factors and set a quick cut-off: Social Impairment ≥ 26 flags autism with a large share accuracy. This extends the 1988 finding by giving you a faster score to watch.
Parks (1983) had warned that early autism scales lacked good validity. The 1988 study answers that worry by showing CARS can separate autistic and non-autistic teens.
Gabriels et al. (2001) validated the M-CHAT for toddlers. Together these papers show CARS works across ages, while M-CHAT covers the youngest group.
Why it matters
If you assess teens for autism, you can keep using CARS. Just skip the one item that lowers the total. This saves you from buying another tool and keeps your data consistent across age groups.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Remove item 12 when scoring CARS for any teen and use the Social Impairment factor ≥ 26 as your quick red-flag check.
02At a glance
03Original abstract
The validity and reliability of the Childhood Autism Rating Scale (CARS) for autistic adolescents was tested. In the first study, CARS scores for autistic children and adolescents (matched on nonverbal IQ, sex, and ethnicity) were compared. In the second study, a group of nonautistic, handicapped adolescents were administered the CARS and these scores were compared with those of a group of autistic adolescents (matched on age, nonverbal IQ, sex, and ethnicity). The CARS clearly discriminated the two adolescent groups, suggesting that the scale may be an adequate measure of autism in adolescence. Although the CARS total score did not discriminate the younger from older autistic subjects, some interesting age-related differences emerged for specific items. A recommendation is made for elimination of one of the items on the CARS that negatively correlates with the CARS total score for both autistic groups.
Journal of autism and developmental disorders, 1988 · doi:10.1007/BF02212193