Defining autism subgroups: a taxometric solution.

Sajith et al. (2008) ran taxometric math on autism data. They asked: do autism traits form smooth slopes or clear steps?

The team looked at social-communication scores, IQ, and daily-living skills. They used large public data sets of people already diagnosed with ASD.

Social-communication and IQ fell into two separate lumps, not a single hill. Daily-living skills stayed smooth.

The authors say ASD is not one long line from mild to severe. Instead, it has subgroups you can sort by social-communication level and IQ.

Shuster et al. (2014) backs the idea. Their review of 36 studies also shows two solid chunks: social/communication and restricted/repetitive behaviors.

Morales-Hidalgo et al. (2018) seems to disagree. In a regular school sample they saw only smooth, continuous traits. The clash fades when you notice G et al. used clinic data. Kids in clinics often have sharper gaps than kids in classrooms.

Qiao et al. (2025) extends the claim. Brain scans on two ASD groups matched the IQ split G et al. found, giving the subgroups a neural footprint.

Stop treating every client as a point on one sliding scale. Look for the social-communication plus IQ profile instead. Pick goals, visuals, and peer models that fit that profile. When you write reports, name the subgroup so the next clinician starts with the same picture.