Assessment & Research

Children on the borderlands of autism: differential characteristics in social, imaginative, communicative and repetitive behaviour domains.

Barrett et al. (2004) · Autism : the international journal of research and practice 2004
★ The Verdict

Joint attention, flexible play, and social use of language separate autism from language disorder better than word counts.

✓ Read this if BCBAs assessing preschool or early-elementary children with unclear diagnoses.
✗ Skip if Clinicians only treating older fluent speakers with no play concerns.

01Research in Context

01

What this study did

Barrett et al. (2004) watched 4- to 7-year-old children with autism and children with language disorder. Both groups talked at the same level on formal tests.

The team scored joint attention, pretend play, everyday language use, and repetitive actions during play sessions. They wanted clear flags that separate autism from pure language delay.

02

What they found

Autistic children shared eye gaze and toys far less often. Their play was simpler and they showed more hand flapping, lining up toys, or other repeated moves.

Even though both groups used the same number of words, autistic kids had trouble using words to ask, show, or joke. These gaps gave a clear picture of autism apart from language disorder.

03

How this fits with other research

Yin et al. (2026) later saw the same play limits in Mandarin-speaking preschoolers, so the play gap holds across languages.

Kaçar Kütükçü et al. (2026) found a twist: Turkish-speaking autistic children actually scored lower than language-disordered peers on grammar and word meaning. The two studies seem opposite, but Suzanne looked at how kids use words socially while Dilber looked at grammar rules. Both can be true; social use and grammar are different skills.

Chakraborty et al. (2021) added stomach trouble to the story. In 176 autistic children, worse constipation or food selectivity went hand in hand with more repetitive movements. Suzanne first flagged the movements; Payal showed one reason they can spike.

04

Why it matters

If a child talks but rarely shows toys, shares looks, or varies play, think autism first even when vocabulary looks fine. Track GI signs when repetitive behaviors jump; treating constipation or food limits may calm the loops. Use typical peer play levels as your teaching target and embed joint-attention drills into every fun activity.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Start each session with a 2-minute joint-attention probe: have the child show, give, or point to share interest; score yes or no and teach missing acts before language drills.

02At a glance

Intervention
not applicable
Design
quasi experimental
Sample size
37
Population
autism spectrum disorder, mixed clinical
Finding
positive

03Original abstract

A sample of 37 children aged 4-7 years who all showed some autistic features was investigated. Children with a primary diagnosis of autism were compared with those diagnosed with a language disorder, on behaviours within four domains; social behaviour, imaginative activities, repetitive behaviour and communication. The aim was to identify potentially differentiating features of the two groups using observational ratings and questionnaire measures provided by parents and teachers. Information on participants' intelligence and language skills was also collected. The children with autism showed greater deficits in joint attention, functional play and pragmatic language, and engaged in more repetitive behaviours, than the language disordered children. However, the groups did not differ significantly on formally assessed language skills. A cluster analysis produced three groups of children varying in level of functioning and parent-rated behaviours. The results are informative for clinicians dealing with the challenge of differential diagnosis.

Autism : the international journal of research and practice, 2004 · doi:10.1177/1362361304040640