The development and functional control of reading-comprehension behavior.

Three kids with autism learned to pick the picture that matched a short story. The teacher first read the story, then asked, "What happened?" and gave candy for the right picture.

The team used an ABAB design. They turned the candy on and off to prove it was the candy driving the learning.

All three kids quickly hit a large share correct choices while candy was available. When candy stopped, accuracy stayed high as long as an adult or peer stayed in the room. Take the person away and scores dropped.

In other words, social presence—not more candy—kept the skill alive.

Hackenberg (2018) shows candy is just one type of backup reinforcer in a token system. The 1976 study proves even simple candy can build new academic skills before you fade to social reward.

Newman et al. (2021) found arbitrary edibles can reduce escape behavior without extinction. Together these papers tell us edible rewards work for both building skills and cutting problem behavior, but you must test each learner.

Renne et al. (1976) looked at which part of a stimulus gains control. Their lab work reminds us that what the learner pays attention to—visual or auditory—depends on the contingency, just like here the contingency shifted from candy to social praise.

You can teach reading comprehension with brief edible rewards, then maintain it with simple social presence. Try fading candy once accuracy hits a large share for two days; keep an adult or peer at the table and track if scores hold. If they dip, add back brief praise or tokens, then thin again. This saves you from endless edible bowls and moves the learner toward natural classroom conditions.