Constructs assessed by the GARS-2: factor analysis of data from the standardization sample.
GARS-2’s subscales fall apart under math checks, so lean on newer tools or the revised GARS-3.
01Research in Context
What this study did
The authors ran a factor analysis on the GARS-2 standardization sample. They wanted to see if the test’s three published subscales really hang together.
The sample came from the manual’s own data set. This let them check whether the structure the test claims actually shows up in the numbers.
What they found
The three subscales did not appear. The items clumped in ways that did not match the manual.
Because the structure failed, the Autism Index loses meaning. The scale may not measure what it says it measures.
How this fits with other research
Lecavalier (2005) saw the same problem in the first GARS. Items scattered and sensitivity was low. The new edition kept the flaw.
Yang et al. (2026) looks like a contradiction. Their Chinese GARS-3 shows strong numbers and 86–89 % accuracy. The difference is version: GARS-3 rewrote items and norms, so it supersedes the 2010 findings.
Sutton et al. (2022) extends the story. Using GARS-3, they found boys score higher on repetitive items while girls only spike on fear. This warns us that even a better scale still needs sex-aware cut points.
Why it matters
If you still keep GARS-2 in a drawer, treat the Autism Index as a rough flag, not a diagnosis. Pair it with ADOS-2 or CARS-2 and document the limitation in your report. When you upgrade to GARS-3, re-check your own data; the new norms may change who you flag.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pull one old GARS-2 file, re-score the Autism Index, and add a note that the number is exploratory only.
02At a glance
03Original abstract
The GARS-2 (Gilliam 2006) is widely used for autism screening; however, the validity of its three conceptually-derived subscales has not been evaluated. In this study, exploratory (EFA) and confirmatory (CFA) factor analyses of data from matched subgroups of the standardization sample did not support the GARS-2 subscale structure. EFAs identified four factors in the model development subgroup (n = 496): stereotyped/repetitive behavior, stereotyped/idiosyncratic language, word use problems, and social impairment. CFAs supported this model in the replication subgroup (n = 494). Findings suggest that the GARS-2 subscales have limited clinical utility and that the Autism Index should be interpreted with caution. Although more research is needed, the clinical utility of the four empirically-derived scales may be limited by factors related to item content and test development procedures.
Journal of autism and developmental disorders, 2010 · doi:10.1007/s10803-010-0967-1