The effects of reinforcement magnitude on skill acquisition for children with autism.

The team ran discrete-trial lessons with children with autism. Each child got the same lesson three ways: big edible reward, small edible reward, or praise only. Sessions rotated in an alternating-treatments design so the child never knew which pay-off was coming next.

Every child picked the big edible when asked, but bigger pay-offs did not speed up learning. Acquisition stayed about the same across large, small, and praise-only blocks. Preference and performance did not match.

Voss et al. (2019) extends the same idea to a new skill. They also used large versus small edibles and saw the opposite: bigger rewards did boost varied communication for their kids. The difference is the target skill—Cox et al. (2015) looked at how fast a brand-new response was learned, while Voss et al. (2019) looked at how many different ways the child could answer.

Northgrave et al. (2019) conceptually replicates the setup but swaps the variable. Instead of size, they compared who picked the reward. They found faster learning when the adult, not the child, chose each time. Together the two papers warn us that bigger or child-chosen rewards feel good yet may not teach faster.

Jason et al. (1985) is an earlier cousin. They showed that switching between sensory and edible rewards beat edibles alone before kids got satiated. Cox et al. (2015) narrows the lens to edible size only and finds no gain, reminding us that type of reward may matter more than amount.

Before you load extra candy into your token pouch, run a quick alternating-treatments probe. Track trials to mastery across big, small, and praise-only conditions for the exact skill you are teaching. If speed is equal, you can save food, calories, and time while keeping motivation high.