Self-conceptualisation in autism: Knowing oneself versus knowing self-through-other.

Farley et al. (2010) asked verbally able teens with and without autism about their self-concept.

They used questionnaires and short interviews.

The team looked at four parts of self-concept: feeling unique, feeling the same over time, feeling in control, and seeing yourself through someone else’s eyes.

Both groups felt equally unique and steady across time.

The autism group scored lower on agency — feeling they cause their own actions.

They scored much lower on “self-through-other” — guessing how a friend would describe them.

Dritschel et al. (2010) ran a near-identical teen study the same year and also found weaker self-understanding in autism, so the result is not a one-off.

Smith et al. (2008) looks like a clash — adults with autism kept a clear sense of agency in lab tasks. The gap fades when you see that adults may learn tricks teens have not, and lab tasks differ from self-report forms.

Dunphy-Lelii et al. (2012) also seems opposite: preschoolers with autism passed a mirror self-recognition test. Again, age and task type explain the difference — spotting your face is not the same as rating your own control.

Atherton et al. (2019) lets teens speak for themselves and shows they view perspective-taking as “different,” not broken. This supports targeting agency and other-view lessons without labeling the skill as “defective.”

When you write social-cognitive goals, add items that boost agency language (“I chose,” “I made it happen”) and other-view practice (“What would your teacher say about you?”).

Use role-play where the client first states their own opinion, then predicts a peer’s view, then checks it in real time.

Keep the tone neutral — teens already feel “different”; framing the skill as a difference to explore reduces push-back and shame.