Reliability of the ADI-R: multiple examiners evaluate a single case.

Seven different clinicians gave the same toddler the Autism Diagnostic Interview-Revised. They worked alone and never talked to each other.

The team then compared every score to see how often the raters agreed.

Agreement was high: 94-96 % of item scores matched across the seven raters. Kappa values were 0.80-0.88, showing strong reliability.

Even with only one child, the ADI-R held together well across independent examiners.

Tassé et al. (2013) later repeated the test with 3- to 18-year-olds in Japan and still found solid reliability, so the tool travels beyond toddlers.

de Bildt et al. (2013) looked at 1,204 Dutch children and showed that different ADI-R cut-off sets change sensitivity and specificity. High rater agreement does not guarantee perfect diagnosis; you still need to pick the right algorithm.

Constantino et al. (2003) offered a 15-minute parent scale that correlates about 0.70 with the full ADI-R. The brief form saves time, but the deep interview remains the gold standard when you can spare the hour.

You can trust the ADI-R even when several clinicians interview the same family on different days. Train new team members with this case: have them watch a master interview, then score the recording and compare answers. Aim for the same 94 % match before they fly solo. If you need a quick screen, pair the short SRS first, but keep the full ADI-R in your pocket for the final diagnostic meeting.