Do Parental Interviews for ASD Converge with Clinical Diagnoses? An Empirical Comparison of the 3di and the DISCO in Children with ASD, a Clinically-Referred Group, and Typically Developing Children.

Kris and team asked 75 parents to complete two gold-standard ASD interviews: the 3di and the DISCO-11. The kids were split three ways: 25 had a clinical ASD label, 25 were in assessment but had other issues, and 25 were typically developing.

Each family did both interviews within two weeks. Trained raters scored the records blind to the child’s group. The goal was simple: how often do the two tools give the same yes-or-no answer?

Overall agreement looked okay at 75 %. But when the team dug into the ASD group, only the kids out of 25 hit the ASD cut-off on BOTH tools. The 3di caught social symptoms the DISCO missed, while the DISCO flagged repetitive behaviors the 3di skipped.

In plain numbers: each tool missed a different slice of the autism picture. Using only one interview would have under-identified more than half of the already-diagnosed kids.

Stephenson et al. (2015) meta-analyzed visual-analysis agreement and landed on the same 76 % figure. The echo is striking: whether you’re eye-balling a graph or tallying interview items, one rater or one tool is rarely enough.

Lam et al. (2011) warn that switching how you record behavior can inflate agreement scores. Kris et al. show a parallel trap: switching interview forms can inflate your confidence while still missing kids. Both papers shout, “Check your method before you trust the number.”

Najafichaghabouri et al. (2024) found that small changes in interviewer style swayed kids’ answers. That lines up with Kris’s finding that different question sets (3di vs DISCO) pull out different symptoms. The child isn’t changing; the lens is.

Next time you screen for ASD, plan for two data sources. Give both the 3di and the DISCO, or follow either one with direct observation. Document which symptom domains each tool covers so you don’t unknowingly drop social-communication or repetitive-behavior evidence. Building this dual-check into your intake packet can catch the 1-in-the kids who would otherwise slip through on a single interview.