Reinforcing variability in adolescents with autism.

Five teens with autism played a simple computer game. Each press on one of three keys made a shape appear on screen.

The game gave points only when the teen made a new three-key sequence. Repeats earned nothing. The team ran an ABAB design: baseline, game, back to baseline, then game again.

During game phases, every teen produced far more unique key patterns. Novel sequences jumped from about a large share to over a large share.

When the game stopped, variability dropped back to baseline. Bringing the game back again lifted variability each time.

Chin Wong et al. (2017) extended the idea to talking. They used prompts like “tell me more kinds” and saw more varied answers from a child with autism. Both studies show you can build flexible behavior in the same diagnosis.

Reed (2023) tested college students without autism. A quick light before each trial also raised variability when the task required it. The two papers together say the principle works across ages and diagnoses.

Leon et al. (2023) seems to disagree. They found that unpredictable room changes triggered hitting and screaming in younger kids with autism. The key difference: Dugdale et al. (2000) let the teens control the change and win points for it, while Leon imposed change without warning. Choice versus surprise makes the results line up, not clash.

If a client repeats the same phrase or play action, build a quick game where new moves earn tokens or screen time. Start with short sets and celebrate every novel response. You may see more flexible social and play skills spill into daily life.