Assessment of autism in community settings: discrepancies in classification.

Lancioni et al. (2009) looked at how often community clinicians agree on an autism diagnosis. They reviewed records of preschoolers sent to public clinics for evaluation.

Two independent reviewers checked each file. They scored whether the child met autism criteria and if the report followed best-practice rules like using a second language when needed.

The reviewers agreed on autism yes-or-no only 45 percent of the time. Most reports skipped key steps such as bilingual testing or clear behavior descriptions.

In short, community teams gave different answers about the same child and rarely used the full checklist experts recommend.

Taylor et al. (2017) saw the same problem. Their expert judges watching ADOS videos also agreed only one-third of the time. Together the two papers show poor reliability across both live clinics and filmed exams.

Stadnick et al. (2015) offers a bright spot. When community clinicians stuck to the ADOS script for school-age kids, their labels matched later diagnoses. The difference: Nicole’s teams had extra ADOS training while E’s teams used free-form reports.

Myers et al. (2018) helps explain the mess. Agreement drops when families are low-income or African American, groups that filled many of E’s clinics. The studies stack up to say: method matters and so does family context.

If you rely on outside evaluations, know that a coin-flip could match the odds those reports agree. Ask for the tools used and look for ADOS, ADI-R, or bilingual steps. When reports conflict, schedule your own direct observation and gather parent plus teacher data before you write goals or recommend services.