Behavioral assessment of joint attention: a methodological report.

The team watched 15-minute toy-play sessions between an adult and the toddlers. Half the kids had autism; half were typically developing. They counted two things: how often the child looked when the adult pointed or showed a toy (responding), and how often the child pointed or showed first (initiating). Two coders scored each video frame-by-frame to check if the numbers stayed the same across people.

Kids with autism responded to adult bids only a little less than typical peers. But they started joint-attention bids far less often. The gap was big enough to see with bare eyes. Inter-coder agreement hit 90 %, so the short protocol is reliable.

Liu et al. (2021) later added eye-tracking software and found the same kids also shift gaze late and out of sync. The 2006 hand-coding still caught the problem; the 2021 tech just gives you micro-timing for free. Harrison et al. (2016) then showed the same joint-attention scores link to verbal IQ and daily living skills in over the kids. So the little 15-minute clip predicts real-world function. Mace et al. (1990) looked one step deeper: autistic kids not only start fewer bids, they also show less happy face while doing it. Put together, you now know to teach both the look and the smile.

You can run this protocol with one adult, one basket of toys, and a phone camera. Score it in 20 minutes. Use the numbers as a baseline, write an objective on the ISP, then re-film every month to show growth. No extra gear, no eye-tracker budget needed.