Strategies for conducting research on language in autism.

Tager-Flusberg (2004) wrote a position paper. She told autism language researchers to stop matching autistic kids to non-autistic controls.

Instead, she said, map language types inside autism. Find subgroups that share the same language profile.

The paper did not test kids. It argued that matched-group designs hide autism’s true language range.

Helen claimed that only within-group phenotyping can show why some autistic children speak, some script, and some stay non-verbal.

Charman (2004), published the same year, lists the exact problems Helen attacks. Tony still gives tips for better matching, while Helen says drop matching entirely. The two papers feel opposite but share the same complaint.

Maes et al. (2023) and Journal et al. (2024) later did what Helen asked. They clustered preschoolers with autism on tiny vocal details and on early social gestures. Both teams found clear sub-groups that predict different language paths.

Wehman et al. (1989) is the old style Helen criticises. That study tracked autistic and dysphasic boys matched on age, then compared their later language. Helen would say the design washed out autism-only patterns.

Stop spending hours hunting for a ‘perfect’ language-matched control. Use that time to profile the child in front of you. Note echoing, jargon, gesture use, and vocal quality. Pick intervention goals that fit that profile, not goals that fit a matched peer.